Free beta access. 5 free AI tokens each week. Optional top-ups available. See what is included
Open evidence library

Search the archive.

Current and older evidence is free during beta. Sign in for unlimited card access and full source downloads.

Matching cards
2022 MNDI Affirmative
AI Ethics › Aff › Extend – Inherency - DOD

The DOD has not committed enough resources to TEVV, despite support for AI

Flournoy, Haines et al 2020 - former Under Secretary of Defense for Policy and Director of National Intelligence [Michèle and Avril, Gabrielle Cheftiz Special Assistant to the Under Secretary of Defense for Policy October, “Building Trust through Testing Adapting DOD’s Test & Evaluation, Validation & Verification (TEVV) Enterprise for Machine Learning Systems, including Deep Learning Systems” https://cset. georgetown.edu/wp-content/uploads/Building-Trust-Through-Testing.pdf Acc 6/23/22 JZ]

Machine learning applications can increase the speed and quality of human decision-making on the battlefield, enable human-machine teaming to maximize performance and minimize the risk to soldiers, and greatly improve the accuracy and speed of analysis that relies on very large data sets. ML can also strengthen the United States’ ability to defend its networks against cyberattacks at machine speeds and has the power to automate critical components of labor-intensive enterprise functions, such as predictive maintenance and personnel management. Advances in AI and machine learning are not the sole province of the United States, however. Indeed, U.S. global leadership in AI remains in doubt in the face of an aggressive Chinese challenge in the field. Numerous DOD and academic reports reflect on the need to invest more in AI T 2 research and development, train and recruit a skilled workforce, and promote an international environment supportive of American AI innovation—all while promoting safety, security, privacy, and ethical development and use. However, far too little attention is placed on the issue of trust, and especially testing, evaluation, verification, and validation (TEVV) of these systems. Building a robust testing and evaluation ecosystem is a critical component of harnessing this technology responsibly, reliably, and urgently. Failure to do so will mean falling behind.

2022 MNDI Negative
Nuclear Human Control Counterplan – 1NC

Text – The United States should increase security cooperation with NATO to require positive human control over all nuclear launch decisions

Counterplan solves the risk of nuclear accidents by requiring human control

Horowitz and Scharre 2021 - Director of the Emerging Capabilities Policy Office in the Office of the Under Secretary of Defense for Policy and  Vice President and Director of Studies at CNAS [Michael, Paul, January 12 2021, “AI and International Stability: Risks and Confidence- Building Measures”, , Acc 6/18/22. M.A.]

Specific Mission-Related CBMs: Nuclear Operations The integration of AI, autonomy, and/or automation into nuclear command-and-control, early warning, and delivery systems poses unique risks to international stability because of the extreme consequences of nuclear accidents or misuse. One option for mitigating these risks could be for nations to set limits on the integration of AI, autonomy, or automation into their nuclear operations. Some U.S. military leaders and official DoD documents have expressed skepticism about integrating uninhabited vehicles into plans surrounding nuclear weapons. The Air Force’s 2013 Remotely Piloted Aircraft (RPA) Vector report proposed that nuclear strike “may not be technically feasible unless safeguards are developed and even then may not be considered for [unmanned aircraft systems] operations.” U.S. Air Force general officers have been publicly skeptical about having uninhabited vehicles armed with nuclear weapons. General Robin Rand stated in 2016, during his time as head of Air Force Global Strike Command, that: “We’re planning on [the B-21] being manned. … I like the man in the loop … very much, particularly as we do the dual-capable mission with nuclear weapons.” Other U.S. military leaders have publicly expressed support for limits on the integration of AI into nuclear command-and-control. In September 2019, Lieutenant General Jack Shanahan, head of the DoD Joint AI Center, said, “You will find no stronger proponent of the integration of AI capabilities writ large into the Department of Defense, but there is one area where I pause, and it has to do with nuclear command and control.” In reaction to the concept of the United States adopting a “dead hand” system to automate nuclear retaliation if national leadership were wiped out, Shanahan said, “My immediate answer is ‘No. We do not.’ … This is the ultimate human decision that needs to be made which is in the area of nuclear command and control.” While the motivation for these statements about limits on the use of autonomy may or may not be strategic stability—bureaucratic factors could also be at play—they are examples of the kinds of limits that nuclear powers could agree to set, unilaterally or collectively, on the integration of AI, autonomy, and automation into their nuclear operations. Nuclear states have a range of options for how to engage with these kinds of risks. On one end of the spectrum are arms control treaties with some degree of verification or transparency measures to ensure mutual trust in adherence to the agreements. On the other end of the spectrum are unilateral transparency measures, which could have varying degrees of concreteness ranging from informal statements from military or civilian leaders along the lines of Shanahan’s and Rand’s statements, all the way to formal declaratory policies. In between are options such as mutual transparency measures, statements of principles, or non-legally binding codes of conduct or other agreements between nuclear states to ensure human control over nuclear weapons and nuclear launch decisions. Even if states that desired these restraints found themselves in a position where others were unwilling to adopt more binding commitments, there may be value in unilateral transparency measures both to reduce the fears of other states and to promulgate norms of responsible state behavior. As with other areas, it is important to consider incentives for defection from an agreement and the extent to which one state’s voluntary limitations depend on verifying others’ compliance with an agreement. If some states, such as the United States, desire strict positive human control over their nuclear weapons and nuclear launch authority for their own reasons, then verifying others’ behavior, while desirable, may not be a necessary precondition to those states adopting their own limits on the use of AI, autonomy, or automation in nuclear operations. Two possible CBMs for AI applications in the nuclear arena involve nuclear weapons states agreeing to strict human control over nuclear launch decisions and ensuring any recoverable delivery vehicles are human-inhabited, to ensure positive human control. Strict Human Control over Nuclear Launch Decisions One CBM for uses of AI in the nuclear arena would involve an agreement by nuclear powers to ensure positive human control over all nuclear launch decisions. This type of agreement would preclude automated “dead hand” systems or any other automatic trigger for the use of nuclear weapons. The benefit of such a CBM would be to reduce the risk of accidental nuclear war. It would preclude a machine malfunction leading directly to the use of nuclear weapons without a human involved in the process. Agreement on positive human control over nuclear launch decisions could also be a mechanism for dialogue with newer nuclear powers, helping generate more transparency over their nuclear launch decisions.

2022 MNDI Negative
--Extend – Military Technology Net Benefit

The counterplan solves for AI safety without revealing TEVV secrets to Russia and China.

Horowitz and Kahn, 2021 – Senior and Research Fellows at the Council on Foreign Relations [Michael and Lauren, The Washington Quarterly 3-19-2021 “Leading in Artificial Intelligence through Confidence Building Measures” TM]

As noted above, one might be opposed to dialogue on AI safety with China and Russia, fearing that it could reveal information that China and Russia would use to make their military applications of AI more reliable, meaning they would be more effective. However, these proposed discussions would not include methods to make algorithms more reliable or divulge details on how TEVV works. Dialogues of that nature might provide China and Russia with a roadmap to make their algorithms more effective. Instead, dialogue would focus on ideas about strategic stability and areas of mutual interest such as reducing the chance of inadvertent escalation. Communicating the importance the United States places on safety and reliability and emphasizing already public information on US safety checks before deploying weapons could also generate the advantages described above.

2022 MNDI Negative
AIA Counterplan – 1NC

Text – The United States should enter into an Autonomous Incidents Agreement with the People's Republic of China.

An Autonomous Incidents Agreement minimizes unintended escalation – precedent shows rules of the road can avoid military accidents

Horowitz and Scharre 2021 - Director of the Emerging Capabilities Policy Office in the Office of the Under Secretary of Defense for Policy and  Vice President and Director of Studies at CNAS [Michael, Paul, January 12 2021, “AI and International Stability: Risks and Confidence- Building Measures”, , Acc 6/18/22. M.A.]

There are inherent risks when autonomous systems with any level of decision-making interact with adversary forces in contested areas. Given the brittleness of algorithms, the deployment of autonomous systems in a crisis situation could increase the risk of accidents and miscalculation. AI-related CBMs could build on Cold War agreements to reduce the risk of accidental escalation, with some modification to account for the new challenges AI-enabled autonomous systems present. States have long used established “rules of the road” to govern the interaction of military forces operating with a high degree of autonomy, such as at naval vessels at sea, and there may be similar value in such a CBM for interactions with AI-enabled autonomous systems. The 1972 Incidents at Sea Agreement and older “rules of the road” such as maritime prize law provide useful historical examples for how nations have managed analogous challenges in the past. Building on these historical examples, states could adopt a modern-day “international autonomous incidents agreement” that focuses on military applications of autonomous systems, especially in the air and maritime environments. Such an agreement could help reduce risks from accidental escalation by autonomous systems, as well as reduce ambiguity about the extent of human intention behind the behavior of autonomous systems. In addition to the Incidents at Sea Agreement, maritime prize law is another useful historical analogy for how states might craft a rule set for autonomous systems’ interactions. Prize law, which first began in the 12th century and evolved more fully among European states in the 15th to 19th centuries, regulated how ships interacted during wartime. Because both warships and privateers, as a practical matter, operated with a high degree of autonomy while at sea, prize law consisted of a set of rules governing acceptable wartime behavior. Rules covered which ships could be attacked, ships’ markings for identification, the use of force, seizure of cargo, and providing for the safety of ships’ crews. Nations face an analogous challenge with autonomous systems as they become increasingly integrated into military forces. Autonomous systems will be operating on their own for some period of time, potentially interacting with assets from other nations, including competitors, and there could be value in establishing internationally agreed upon “rules of the road” for how such systems should interact. The goal of such an agreement, which would not have to be as formal as the Incidents at Sea Agreement, would be to increase predictability and reduce ambiguity about the behavior of autonomous systems. Such an agreement could be legally binding but would not necessarily need to be in order to be useful. It would likely need to be codified in an agreement (or set of agreements), however, so that expectations are clear by all parties. An ideal set of rules would be self-enforcing, such that it is against one’s own interests to violate them. Examples of rules of this kind in warfare include prohibitions against perfidy and giving “no quarter,” violating either of which incentivizes the enemy to engage in counterproductive behavior, such as refusing to recognize surrender or fighting to the bitter end rather than surrendering. An autonomous incidents agreement could also include provisions for information-sharing about potential deployments of autonomous systems in disputed areas and mechanisms for consultation at the military-to-military level to resolve questions that arise (including potentially a hotline to respond to incidents in real time).

2022 MNDI Negative
--Extend – Training Data

Effective AI requires enormous amounts of data inputs, which are less available and more protected in the military sphere.

Goldfarb and Lindsay, 2022 – Chair in AI and health care at the Univ of Toronto, Prof of Cybersecurity at the Georgia Tech [Avi, Jon, 2/25/22, , “Prediction and Judgement: Why Artificial Intelligence Increases the Importance of Humans in War” 6/18/22, LND]

Michael Horowitz describes AI as “the ultimate enabler” for automating decision-making tasks in everything from public administration and commercial business to strategic intelligence and military combat. In 2018, the Department of Defense observed that “AI is poised to transform every industry, and is expected to impact every corner of the Department, spanning operations, training, sustainment, force protection, recruiting, healthcare, and many others.” We would be surprised, however, if AI transformed all these activities to the same degree for all actors who use it. One of the key insights from the literature on the economics of technology is that the complements to a new technology determine its impact. AI, from this perspective, is not a simple substitute for human decision-making. Rapid advances in machine learning have improved statistical prediction, but prediction is only one aspect of decision-making. Two other important elements of decision-making—data and judgment—represent the complements to prediction. Just as cheaper bread expands the market for butter, advances in AI that reduce the costs of prediction are making its complements more valuable. AI prediction models require data, and accurate prediction requires more and better data. Quality data provide plentiful and relevant information without systemic bias. Data-driven machine prediction can efficiently fill in information needed to optimize a given utility function, but the specification of the utility function ultimately relies on human judgment about what exactly should be maximized or minimized. Judgment determines what kinds of patterns and outcomes are meaningful and what is at stake, for whom, and in which contexts. Clear judgments are well specified in advance and agreed upon by relevant stakeholders. When quality data are available and an organization can articulate clear judgments, then AI can improve decision-making. We argue that if AI makes prediction cheaper for military organizations, then data and judgment will become both more valuable and more contested. This argument has two important strategic implications. First, the conditions that have made AI successful in the commercial world—quality data and clear judgment—may not be present, or present to the same degree, for all military tasks. In military terms, judgment encompasses command intentions, rules of engagement, administrative management, and moral leadership. These functions cannot be automated with narrow AI technology. Increasing reliance on AI, therefore, will make human beings even more vital for military power, not less. Second, the importance of data and judgment creates incentives for strategic competitors to improve, protect, and interfere with information systems and command institutions. As a result, conflicts over information will become more salient, and organizational coordination will become more complex. In contrast with assumptions about rapid robot wars and decisive shifts in military advantage, we expect AI-enabled conflict to be characterized by environmental uncertainty, organizational friction, and political controversy. The contestation of AI complements, therefore, is likely to unfold differently than the imagined wars of AI substitutes.

2022 MNDI Negative
--Extend – JAIC Fails

The Plan will fail - lack of JAIC authority makes budgeting impossible

Tarraf, Shelton, Parker 2019 - Senior Information Scientist, Senior Engineer, and physical Scientist at the RAND Corporation [ Danielle C., William, Edward, et al RAND Cooperation, “The Department of Defense Posture for Artificial Intelligence”, , Acc 6/18/22, M.A.]

The JAIC lacks the authorities to carry out its present role. At its core, the JAIC’s overarching mission can be distilled to this: Scale AI and its impact across DoD. This mission and its present scope—as defined by the summary of the DoD AI strategy and the memo establishing the JAIC—are extensive, while the JAIC’s current authorities are limited. In particular, the JAIC is expected to synchronize DoD AI activities and coordinate AI initiatives totaling more than $15 million annually. It is unclear whether the JAIC has any mechanisms for enforcing these directives, because it does not have the authorities to direct investments or to halt programs or activities that are deemed to be misaligned with DoD’s strategy (a fact we learned through multiple DoD interviews). In short, the JAIC does not have directive or budget authorities, and that critically limits its ability to synchronize and coordinate DoD-wide AI activities to enact change. Currently, it can catalogue these activities, but it is unclear how doing so would help scale AI across DoD. Of course, that assumes that what constitutes an AI activity is known. However, it is not currently clear how the determination of what constitutes an AI initiative or activity is made, by whom, and whether that determination is consistent across DoD.8

JAIC fails – lack of long term funding certainty undermines industry support

2022 MNDI Negative
--Extend – Adaptive

No Solvency - TEVV does not pair well with formal methods of research

Haugh, 2018 - project leader for the Institute for Defense Analysis [Brian A., September “The Status of Test, Evaluation, Verification, and Validation (TEV&V) of Autonomous Systems,” , 6/23/22 MD]

3. Potential Approaches to Overcome These Challenges Academic and government researchers have begun work on a number of techniques and methodologies that might help to overcome these core challenges. Current efforts fall into four primary categories: formal methods, cognitive instrumentation, adversarial testing, and run-time monitoring. We discuss each of these below, then consider the question of how well they cover the full set of capability gaps implied by the challenge list. A. Formal Methods 1. Summary Formal methods in software development allow developers to specify certain properties that the software should have, produce the software, and verify that it does have those properties without needing to confirm that empirically by testing for them. Properties to be specified might be things like: Property 1: the weapon cannot fire while turret is still rotating Property 2: the course-of-action selector can never get into an infinite loop There are two approaches to formal methods for autonomous systems verification: (1) formal methods can be used after the fact as an analytic tool to verify some properties of existing software, or more importantly, (2) formal methods can be used as a design and development process that can assure much more about the behavior of the software to be developed. Formal methods of the second kind are most commonly used in the development of complex safety-critical or security-critical systems or for expensive one-time development efforts (e.g., deep space probes). Applying formal methods to complex AI and autonomous systems is a natural extension of this. 2. Limitation Although formal methods can be extremely useful, there are significant constraints on the current state of the art. These include the following: Scalability: There are currently fairly tight bounds on the size of development effort (or state space) that the techniques can be applied to. 3-2 Scope: Not all desired properties can be assured through formal methods, and there may be performance trades associated with achieving assurance. Rigidity: Any change to a system developed using formal methods risks invalidating the assurance proofs, unless the formal methods are reapplied to the new specification. Given that not all desired behaviors can be assured using formal methods, there are also open research questions concerning how to combine formal methods with empirical TEV&V techniques or run-time monitoring strategies.

2022 MNDI Negative
Solvency Resps

17. TEVV cannot solve – normal testing methods don’t include run time monitoring which is key to AI

Haugh, 2018 - project leader for the Institute for Defense Analysis [Brian A., September “The Status of Test, Evaluation, Verification, and Validation (TEV&V) of Autonomous Systems,” , 6/23/22 MD]

D. Run-Time Monitoring Given the difficulty of assuring that a system will not exhibit specific undesired behaviors, a natural thought is to instead monitor the system during operations and intervene when bad behavior is imminent. This approach is already common in engineering practice for safety-critical systems. It is mentioned explicitly in the recent US Unmanned Systems Integrated Roadmap 2017-2042: For the most demanding adaptive and non-deterministic systems, a new approach to traditional TEVV will be needed. For these types of highly complex autonomous systems, an alternate method leveraging a run-time architecture that can constrain the system to a set of allowable, predictable, and recoverable behaviors should be integrated early into the development process. Emergent behaviors from large-scale deployment of interacting autonomous systems poses a difficult challenge. The analysis and test burden would thereby, be shifted to a simpler, more deterministic run-time assurance mechanism. The effort for new approaches to TEVV endeavors to provide a structured argument, supported by evidence, justifying that a 3-6 system is acceptably safe and secure not only through offline tests, but also through reliance on real-time monitoring, prediction, and fail-safe recovery.1 Although this mechanism might indeed be simpler than avoiding unpredictable behaviors in the first place, it is not without its own challenges. In general, any behavior whose dependability cannot be adequately assured through system design and training would need to be monitored, with a robust intervention standing by. This means not only intervening when the system is about to execute some undesired physical action (e.g., one that might risk harm to the system or to humans), but also intervening in any case where the system is making a bad decision or misinterpreting its environment. Detecting such cases and handling them gracefully will not always be easy. Research is required into architectures to support this concept, instrumentation needs and control algorithms to predict and avoid specific failure modes, systematic identification of conditions to be monitored for, robustness against attacks designed to invoke fail-safe behaviors, and so forth. It goes without saying that the fail-safe systems would themselves need to be verified and validated as well.

2022 MNDI Negative
AT No Arms Race

Winning the arms race is key to peace - AI enhances deterrence by improving military operations – it lifts the fog of war, accelerates decision making and saves soldiers’ lives

Nurkin and Konaev, 2022 - senior fellows at the Center for Strategy and Security at the Atlantic Council [Tate and Margarita, 5-25-2022, “Eye to eye in AI: Developing artificial intelligence for national security and defense” Atlantic Council ARD]

AI embodies a significant opportunity for defense policymakers. The ability of AI to process and fuse information, and to distill data into insights that augment decision-making, can lift the “fog of war” in a chaotic, contested environment in which speed is king. AI can also unlock the possibility of new types of attritable and single-use uncrewed systems that can enhance deterrence.2 It can help safeguard the lives of US service members, for example, by powering the navigation software that guides autonomous resupply trucks in conflict zones.3 While humans remain in charge of making the final decision on targeting, AI algorithms are increasingly playing a role in helping intelligence professionals identify and track malicious actors, with the aim of “shortening the kill chain and accelerating the speed of decision-making.”4 AI development and integration are also imperative due to the broader geostrategic context in which the United States operates—particularly the strategic competition with China.5 The People’s Liberation Army (PLA) budget for AI seems to match that of the US military, and the PLA is developing AI technology for a similarly broad set of applications and capabilities, including training and simulation, swarming autonomous systems, and information operations—among many others—all of which could abrogate the US military-technological advantage.6 As US Secretary of Defense Lloyd Austin noted in July 2021, “China’s leaders have made clear they intend to be globally dominant in AI by the year 2030. Beijing already talks about using AI for a range of missions, from surveillance to cyberattacks to autonomous weapons.”7 The United States cannot afford to fall behind China or other competitors.

2022 MNDI Negative
Neg › Logistics AI Solvency Resps

Incremental US AI development speeds up the Arms Race – allies will lose trust, and competitors will see an opportunity

Horowitz, 2018 – Professor of Political Science at UPenn [Michael, September “The Algorithms of August: The AI arms race won't be like previous competitions, and both the United States and China could be left in the dust,” https://foreignpolicy.com/2018/09/12/will-the-united-states-lose-the-artificial-intelligence-arms-race/ 6/18/22 MD]

It's also possible, though unlikely, that AI will propel emerging powers and smaller countries to the forefront of defense innovation while leaving old superpowers behind. Washington's current focus on U.S.-Chinese competition in AI misses an even more important trend. There is a risk that the United States, like many leading powers in the past, could take an excessively cautious approach to the adoption of AI capabilities because it currently feels secure in its conventional military superiority. That could prove to be a dangerous form of complacency, especially if relations between the United States and many of its current and partners continue to fray over time. Faced with a less reliable United States, shunned NATO partners would, for example, have even more incentives to invest in alternatives, such as experimenting more with how AI can bolster their capabilities in a world without clear superpower leadership. If these countries decide to strike out on their own, while China and Russia continue to invest in capabilities with the explicit goal of disrupting U.S. military superiority, and parts of the U.S. tech industry remain reluctant to work with the Defense Department, the U.S. military could even find itself in a position it has not faced for more than 75 years: playing catch-up when it comes to deploying cutting-edge technology on the battlefield.

2022 MNDI Negative
Neg › Logistics AI Solvency Resps

AI assisted Logistics break down in practice – the political and bureaucratic conflicts and fog of war make “low hanging fruit” an illusion.

Goldfarb and Lindsay, 2022 – Chair in AI and health care at the Univ of Toronto, Prof of Cybersecurity at the Georgia Tech [Avi, Jon, 2/25/22, , “Prediction and Judgement: Why Artificial Intelligence Increases the Importance of Humans in War” 6/18/22, LND]

We expect automation to improve the efficiency and scale of routinized activities that entail filling in missing information, measuring technical performance, tracking personnel, and anticipating future needs. Indeed, AI may enhance many routine tasks associated with developing budgets, recruiting and training personnel, identifying leadership potential, scheduling unit rosters, designing and procuring weapon systems, planning and evaluating exercises, caring for the morale and welfare of service members and their families, and providing health care to service members and veterans. At the same time, it is important to recognize that seemingly trivial procedures can become politicized when budgets and authorities are implicated. Even in the absence of parochialism, the complexity of administrative systems introduces interpretive challenges for personnel. These internal frictions undermine the conditions for successful administrative automation. Logistics supply chains may also be good candidates for automation. Indeed, firms like DHL and FedEx have leveraged AI to streamline their delivery networks. Standardized parts, consumption rates, repetitive transactions, and preventive maintenance schedules generate abundant data about defined tasks. Using historical performance data, predictive maintenance systems can monitor consumption rates and automatically order replacement parts before a weapon or platform breaks. For example, one U.S. Air Force system uses a predictive algorithm to decide when mechanics should perform an inspection, which allows them to tailor the maintenance and repairs for individual aircraft rather than adhere to generic schedules. But we contend that the prediction of supply and demand for just-in-time delivery will be more difficult in war. While bureaucrats may be insulated from the turmoil of the battlefield, supply lines are more exposed. The enemy can interdict or sabotage logistics. As wartime attrition consumes spare parts, units may squabble about which ones should be resupplied. Friendly units may resort to using platforms and parts in unconventional ways. All this turbulence will cause predictions to fail, which essentially shifts AI into the category of premature automation, discussed below. The classical military solution to such problems is to stockpile an excess of supplies, precisely because wartime consumption is so hard to predict. If organizations eliminate slack resources with AI systems in pursuit of efficiency, however, then they may sacrifice effectiveness when systems encounter unforeseen circumstances. In sum, we expect AI to be most useful for automating routine tasks that are bureaucratically insulated from battlefield turbulence. Administration and logistics tasks that are repetitious and standardized are more likely to have both quality data and clear goals. Humans still provide judgment to define those clear goals, but this happens in advance. Although these conditions are ideal for automation, they can be elusive in practice, especially if there are contested resources and personnel decisions. As a result, even the low-hanging fruit applications will often fall into the other three categories in , particularly human-machine teaming.

2022 MNDI Negative
Neg › Cybernetics K Links

Requiring Human Control is a trap – it stops our resistance to the Distancing and Elevation of modern warfare.

Kalpozos, 2020 - Prof of Law at Harvard [Ioannis 3-16-2020, , Leiden Journal of International Law, “Double elevation: Autonomous weapons and the search for an irreducible law of war", https://papers.ssrn.com/sol3/papers.cfm?abstract_id=3545332 accessed on 6-18-2022]

‘Fully autonomous systems’ then, in relation to war-fighting, are systems that, once deployed, are able to adapt, receive, and process feedback, and display a level of functional autonomy that effectively does not distinguish them from human decision makers.24 If anything, in fact – and that is really the point – such systems may be capable of a higher level of tactical or even strategic decision-making (cognitive) and war-fighting (kinetic) capacity. This seems to be a shared understanding among states and NGOs, otherwise holding different positions in the autonomous weapons debate. Accordingly, the US Department of Defense refers to ‘[a] weapons system that, once activated, can select and engage targets without further intervention by a human operator’.25 Human Rights Watch, while setting out its position against autonomous weapons systems, defines them as ‘[r]obots that are capable of selecting targets and delivering force without any human input or interaction’.26 Other states27 and organizations28 provide similar definitions. These functions of autonomy, and their escalation, entail both physical and cognitive distancing from human agents in the overall process of targeting. Although the concept and image of autonomy dominates the discourse in a way that influences, as we will see in the next section, much of the scientific research and development, it only partially describes technological escalation. In fact, the artificial intelligence of war-fighting complements increasing autonomy with what has been called ‘merged heteronomy’. Increasing cognitive autonomy, and distance, coexists with increasingly close physical proximity. This serves pragmatic aims: Technological limitations in the fragmented development of different aspects of autonomy, the continuing necessity of human input, and the fragility of human/machine networks mean that the individual’s continuing presence in the loop remains an operational necessity. This also means that a full-on confrontation with social and political resistance to the reality of distinct killer robots is placed in abeyance. Continuous, if vague, assurances of the human remaining in the loop and retaining meaningful control are thereby facilitated. And yet, the abeyance is a trap and our presence ‘in the loop’ is no guarantee. To the extent that ‘full autonomy’ is understood to require the separate physical existence of an, often anthropomorphized, robot other, it obscures the crucial role that increasing cognitive autonomy plays in a nominally heteronomous decision-making process. Cognitively, as well as physically and kinetically, humans and machines become decreasingly separate, less and less other. Their understanding of the rules, the nomos, is merging, as is their physical existence. Increasing autonomy and merged heteronomy serve the same purpose, the same teleology of mechanization. ‘The loop’, itself, is changing, increasingly relying on artificial intelligence.

2022 MNDI Negative
Neg › Cybernetics K Links

The Affirmative relies on dehumanizing assumptions – the argument that laws can be established to prevent AI is based on the same assumption that LAWs can perfect human warfare. Both rely on enlightened rationalist ideas of progress, ignoring how this perfection numbs us to dehumanizing violence.

Kalpozos, 2020 - Prof of Law at Harvard [Ioannis 3-16-2020, , Leiden Journal of International Law, “Double elevation: Autonomous weapons and the search for an irreducible law of war", https://papers.ssrn.com/sol3/papers.cfm?abstract_id=3545332 accessed on 6-18-2022]

This article intervenes in a rather crowded debate. Why is there such a flurry of current interest in autonomous weapons and their legal regulation? A visceral reaction to the idea of mechanized killing, combined with the perception of the acceleration of artificial intelligence research,12 may be a sufficient answer. The anticipation of incoming practical problems to-be-solved, questions posed by powerful interested parties, may also be an incentive for lawyers to place their analyses in the marketplace of ideas. But I believe there is more to it. The present is the time to imagine the future, especially when one is both propelled by and unmoored from the past. The mix of artificial intelligence, war, and law draws some of its headiness from a specific historical moment – a culmination of and a departure from enlightenment rationalism; the apex of progress and the moment when we fear it will get out of hand. ‘The end of the end of history’13 marks a daunting beginning. The law of war carries this tension, with respect to both the articulation of rules and their enforcement. The legal institution of war rests on the survival of a soldier’s individual sense of humanity at the time when his life is laid out for the collective. So does individual (criminal) liability: without that assumption – without that fiction – it makes no sense. I argue that the increasing mechanization of warfare pursues the creation of distance from our enemies and from ourselves and reduces the knowledge and intelligence of the law and its application through individual judgement. I further argue that legal research on new weapons technology should focus less on questions of compatibility between given legal rules and the algorithmic and kinetic features of new weapons, as most current scholarship does, and more on an understanding of law as non-reducible to algorithmic engineering. I start, in Section 2, by reviewing the state of the art-in-the-making. Rather than providing an exhaustive taxonomy I aim to highlight the teleological nature of artificial intelligence research and the industry’s investment in the dialectic of increasing machine autonomy and human/machine merging, or ‘merged heteronomy’. I then see, in Section 3, this relationship between technology and war in some historical perspective, with a view to discerning functions relevant to law. I argue that the role of technology in war entails a double elevation: above one’s enemy and above oneself. The elevation above one’s enemy, discussed in Section 3.1, serves a both offensive and defensive impetus and aspires to both spatial and moral/civilizational distance. The elevation above oneself, discussed in Section 3.2, is for self-perfection. It is often associated with a certain understanding of Cartesian dualism and a belief in rational improvement that may see humanity as the cause of inhumanity and de-humanization as our best chance for humanization. It seeks to mechanize judgement and therefore establish a distance from human failings. It is served by war as governance from a distance and by the increasing physical and cognitive merging of humans and machines. Both physically and, to some extent, in moral and civilizational terms, technology and automation promise such improvement through establishing a distance from the human – our human enemy and our human self. The establishment of this distance entails a decreasing role for human judgement and the weakening of responsibility for such judgement. I further argue, in Section 3.3, that law, or certain strands of mainstream jurisprudence, are complicit in such mechanization, to the extent that law is treated as logic even conceivably reducible to algorithm. And that the reaction against this process of distancing and de-humanization, to the extent that it idealizes the proscriptive or regulatory role of law, is bound to disappoint.

2022 MNDI Negative
Neg › Cybernetics K Links

Autonomous weapons reinforce the idea of perfection through technology that will elevate us above our enemies. This elevation is always accompanied by Imperial violence.

Kalpozos, 2020 - Prof of Law at Harvard [Ioannis 3-16-2020, , Leiden Journal of International Law, “Double elevation: Autonomous weapons and the search for an irreducible law of war", https://papers.ssrn.com/sol3/papers.cfm?abstract_id=3545332 accessed on 6-18-2022]

3. Double elevation and the distancing of judgement Technology is not neutral.68 Assuming the neutrality of technology – and attaching to it the assumed neutrality of law – precludes any critical understanding of either. Technology, its development and use, reflects both theoretical and practical commitments: it ‘is covert philosophy’.69 It exists in a relationship of co-production with culture, politics, and law.70 As Paul Edwards put it in a seminal study of Cold War weapons technology, ‘we can make sense of the history of computers as tools only when we simultaneously grasp their roles as metaphors in … the period’s … science, politics and culture’.71 In this section, I argue that the promise of technological progress and automation in war, as in general, is a promise of civilization, a promise of improvement. It entails a double elevation: above one’s enemy and above one’s self. At the centre of it there is a paradoxical assumption, namely that the non-human can be more humane than the human. The elevation above one’s enemy combines military distance with a perception of civilizational and moral superiority. The elevation above oneself aims at creating a distance from human features perceived as weak or unreliable. Both full autonomy and merged heteronomy require the increasing mechanization of human judgement. What we have learned to understand as the civilization of war-fighting rests on and pursues its mechanization. 3.1. Rising above one’s enemy Technological distancing aims at developing asymmetry and invulnerability and elevating oneself above one’s enemy in both strictly speaking military and broader civilizational terms. The latter type of elevation allows not simply a geographical distance but also a moral distance with significant consequences for the role of law and judgement in killing. Military technology is central to early imperialist expansion72 and its concomitant civilizational pretension, culminating in the steep military and moral asymmetry achieved in nineteenth century colonial warfare. Churchill’s description in the context of the Boer war of the British infantry ‘steadily and solidly’ firing against the Sudanese Dervishes in ‘the most signal triumph ever gained by the arms of science over barbarians’, while ‘the mere physical act became tedious’,73 is illustrative. Technology allows military superiority, guaranteeing the physical safety and invulnerability of one’s forces; the asymmetry achieved reflects an already assumed civilizational distance which allows a moral dissociation from the act of killing, expressed in the ennui of physical exertion; the civilization of the technologically advanced party is enforced.74 The role of military technology in the elevation above one’s enemy is most closely associated with the growth of air power and the aspirations of invulnerability associated with it. Air power, especially in situations of colonial asymmetry, constituted a relationship of vertical distance, allowing the surveillance and policing of one’s inferior enemy, both at initial conquest and through the protracted practice of colonial administration and pacification.75 That colonial relationship achieved new technological heights in the context of the Cold War. Towards the end of the 1960s, the Vietnam impasse pushed for the assertion of asymmetry through the development of an automated battlefield to improve targeting capacity and protect American soldiers. Operation Igloo White attempted the surveillance of the Ho Chi Minh Trail in Laos through the use of camouflaged sensors designed to detect different types of human activity, including body heat, vehicle noise or the smell of human urine,76 or sweat.77 When picked up, such activities appeared on the screens in the headquarters’ terminals in Thailand and fed into the targeting system of military aircraft. A ‘kill box’78 was constructed and targeted. The operation’s centralized, computerized, automated method of ‘interdiction’ relied on an active global defence and aspirations for the full automation of the battlefield. These are set out by General William Westmoreland, the Chief of Staff of the US Army at the time, in a tenor strongly evocative of our present debate: On the battlefield of the future, enemy forces will be located, tracked, and targeted almost instantaneously through the use of data links, computer assisted intelligence evaluation, and automated fire control. … I see battlefields on which we can destroy anything we locate through instant communications and the almost instantaneous application of highly lethal firepower. … [A]n improved communicative system … would permit commanders to be continually aware of the entire battlefield panorama down to squad and platoon level … I am confident the American people expect this country to take full advantage of its technology - to welcome and applaud the developments that will replace wherever possible the man with the machine … With cooperative effort, no more than 10 years should separate us from the automated battlefield.79 As it turns out, Operation Igloo White was a complete failure.80 And yet the technological ambition remained. In 1973 the New Scientist echoed General Westmoreland’s technological/military optimism. There was ‘at present, great interest in the development of remotely piloted vehicles (RPV’s) for missions such as reconnaissance, electronic warfare, ground attack and air-to-air combat’.81 Increasingly, to these purposes was added another: targeted assassination. The ambition of the precision of a self-sustaining intelligence/targeting loop in drone warfare illustrates the confluence of offensive and defensive imperatives in elevating oneself above one’s enemy. Markus Gunneflo has shown how the practice of and legal justification for targeted killings was developed in Israeli and US policy as a means of constitutional protection of the citizens to be distinguished from unlawful assassination.82 In such ‘active defence’, especially when exercised globally, we see the merging of the offensive distance of air power, seeking to impose a vertical relationship of war, and the defensive distance of integrated human/machine surveillance systems. This vision is reflected in the prioritization in the 1990s of drone research.83 Drones, both the surveillance and the targeting kind, have been seen as symbolizing a ‘change of paradigm’ in the conduct of war.84 They, however, follow the trajectory discussed – that of achieving an elevation above one’s enemy, associated with geographical distancing and the moral/civilizational distance associated with governing through war from above.85 The present ambition, of both escalated weapon emancipation and human/machine merging, follows that same path. However autonomous, further distancing remains the goal.86 This is not a paradigm change.87 However, to the extent that there is a rapid acceleration of technological development we could, perhaps, refer to an ‘avalanche’: ‘when conditions are ripe, individual events, even small ones, can trigger a massive, downward rush’.88 This metaphor may serve to describe a well-established trajectory combined with the feeling that things may be spiralling out of control. From colonial asymmetry to the post-Cold War fighting of ‘terror’, the elevation above one’s enemy through weapons technology guarantees physical and moral distance; it also denotes, and imposes, the pretention of a higher civilization. As we will see, the promise of precision, professionalization, optimization of decision making – with humans involved, but assisted by technology – underlines another kind of elevation: one that supposedly saves humans from themselves.

2022 MNDI Negative
Neg › Cybernetics K Links

Focusing on Human Control ignores the distancing caused by merged human/machine intelligence – it hides the violence in Human Control, and the inability of regulation to resolve our fears.

Kalpozos, 2020 - Prof of Law at Harvard [Ioannis 3-16-2020, , Leiden Journal of International Law, “Double elevation: Autonomous weapons and the search for an irreducible law of war", https://papers.ssrn.com/sol3/papers.cfm?abstract_id=3545332 accessed on 6-18-2022]

The major stance in opposition to technological escalation towards full autonomy is centring on the taboo of delegating life/death decisions to machines and the innate inability of machines to properly apply law. A public expression of this position is that of the Campaign to Ban Killer Robots.144 The campaign is notable in resisting both superficial techno-optimism and dangerously instrumental pragmatism. The potential power of mobilization of public opposition notwithstanding, the position presents some analytical, strategic and conceptual shortcomings. Firstly, the Campaign’s approach starts from an assumption of a fundamental ‘change of paradigm’.145 This does not always appreciate the continuing role of technology in war and its function in the double elevation described here. Indeed, I have argued that the escalation towards full autonomy and merged heteronomy represents a continuation of an existing trajectory of distancing through the elevation above one’s enemy and above oneself, albeit with a perhaps justified presentiment of an ‘avalanche’, a violent acceleration of pace, out of control. Secondly, present practice does not suggest that a prospect of successful imposition of a ban or moratorium is realistic. This is due both to the anticipated military advantage and the inscription of this process in socioeconomic structures and expectations, as reflected in the evidence of enthusiastic investment in the acceleration of this trajectory. Instead, as the analysis above has suggested, while the absolute of full autonomy (combining kinetic and cognitive elements) is kept at bay, the ground is constantly prepared. Thirdly, the primary focus on preventing full autonomy or, inevitably anthropomorphized, ‘killer robots’ is in danger of missing the target. Autonomy is complemented by increasingly merged heteronomy. As discussed in Section 2.1 above, merged heteronomy addresses the logistical limitations of spatially spread human/machine networks. Crucially, while presenting itself as respecting the moral taboo of life-and-death delegation, merged heteronomy advances the mechanization of judgement in pursuit of double elevation. To the extent that the maintenance of ‘meaningful human control’ is primarily focused on ‘keeping humans in the loop’, it is in danger of ignoring the gradual change in the nature and function of that very loop. As reliance on artificial intelligence increases, it is humans who are becoming the ‘killer robots’. Finally, to the extent that the position relies on the incompatibility of autonomous weapons with international humanitarian law, it may be vulnerable to the complicity, discussed in Section 3.3, of certain strands of legal thinking with an understanding of knowledge as ‘a large store of neutral data’146 and the promise of the piece-meal resolution of technical legal problems. It also allows one’s intuitive angst to be assuaged by promises of, or indeed steps towards, the panacea of global regulation.147 As important as such regulation may be in structuring the ambitions of both state and private actors, it would not, per se, address the most fundamental dangers of the mechanization of judgement. Law will neither ban nor regulate away what causes our angst. To the contrary, it may be adapted to serve mechanized judgement. If we are to oppose double elevation and the mechanization of judgement, and hope to use law to this effect, we need legal thinking to serve this purpose. Otherwise, all we can do is surrender to the stance of agnostic abeyance, until the code is engineered.

2022 MNDI Negative
Neg › --Extend – No Definition of Autonomous Weapons

No solvency-- we cannot develop proper AI without first defining AI, which we have not yet done

Roff, 2019 - research fellow at the Centre for the Future of Intelligence at Cambridge [Heather M., 29 April “The frame problem: The AI “arms race” isn’t one,” , 6/22/22, MD]

There needs to be a change in thinking about AI. Those dealing with AI must insist on greater clarity about its definition. If policy makers and other leaders are not clear about what the term means and entails, they cannot possibly formulate best practices and governance mechanisms. It would help matters if artificial intelligence discussions were framed in an “AI +” framework, because in many cases, AI is merely a tool included in a system involving other functions or capabilities. The news media should stop framing the global artificial intelligence competition as an “arms race.” This misrepresents the competition going on among countries. The policy community needs a clear-eyed appraisal of AI’s capabilities and limitations. Without that orientation, those who hope to steer research and development in positive directions will create more problems than they solve.

2022 MNDI Negative
Neg › --Extend – No Definition of Autonomous Weapons

There is no agreed upon definition of Autonomous Weapons

Scharre, 2016 -- Vice President and Director of Studies at CNAS [Paul, Feb 2016, Center for New American Security, “Autonomous weapons and operational risk”, https://s3.us-east-1.amazonaws.com/files.cnas.org/documents/CNAS_Autonomous-weapons-operational-risk.pdf?mtime=20160906080515&focal=none, Acc 6/21/22, M. A.]

The term autonomous is sometimes reserved for systems that exhibit some degree of learning, adaptation, or evolutionary behavior. Others, however, might use the term “autonomous” to refer to complex rule-based systems that exhibit goal-oriented behavior, systems that some might call “automated.” Examples of learning systems include robots that teach themselves how to move around their environment or the Nest “learning thermostat.”9 • Finally, we sometimes refer to autonomous systems that are capable of human-level cognitive tasks, at least for narrow problems, as “intelligent.”

2022 MNDI Negative
Neg › --Extend – Human Control Fails

Human control cannot solve all risks – delay in response times increases the chance of mistakes

Scharre, 2016 -- Vice President and Director of Studies at CNAS [Paul, Feb 2016, Center for New American Security, “Autonomous weapons and operational risk”, https://s3.us-east-1.amazonaws.com/files.cnas.org/documents/CNAS_Autonomous-weapons-operational-risk.pdf?mtime=20160906080515&focal=none, Acc 6/21/22, M. A.]

In a semi-autonomous system, after each task the machine stops and waits for human approval before continuing. Under this control type, the human operator has the ability to observe the machine’s actions in the environment and confirm that the behavior is appropriate before continuing.7 In supervised autonomous operation, in principle the human operator has the ability to intervene if necessary. In practice, there is likely to be some time delay between when a failure occurs and when the human is able to actually correct the behavior of the system. This could occur because of a time delay in communications or because it may take the human some period of time to understand that the system is performing inappropriately and determine the appropriate corrective action. For fully autonomous systems, the human operator lacks the ability to observe the autonomous system’s behavior and undertake corrective action in sufficient time if the system fails to perform appropriately. Presumably at some point in time the human operator will become aware of how the system performed. For example, a household thermostat is operating “fully autonomously” while one is away from the home. Once one returns home, one discovers whether the thermostat was performing as one intended or not.8 Thus a key element of risk in autonomous systems is the time between when a system begins failing (performing in a manner other than what the human operator intended) and when the human operator can undertake corrective action. Even in fully autonomous systems, presumably the system ceases operation at a certain point in time once the task is complete and the results of its actions can be observed. The damage potential of a system depends on its inherent hazard and the time from failure to corrective action When it comes to risk, we are concerned with an autonomous system’s damage potential. Damage potential is the amount of damage an autonomous system could do, if it failed to perform appropriately, before a human operator could take corrective action. Damage potential depends upon the inherent hazard of the system—the type of task being performed and the environment in which it is operating—as well as the control type. For supervised autonomous systems, the speed of the system’s operation and any potential time delays also are significant factors. A system that in principle has a human on the loop to intervene in the event of system failure might in practice still have a high damage potential if the system performs tasks much more rapidly than human operators can react.

2022 MNDI Negative
Neg › --Extend – Human Control Fails

Human control cannot solve for instability – adversaries can never know if the system’s behavior followed human intentions or AI

Horowitz and Scharre 2021 - Director of the Emerging Capabilities Policy Office in the Office of the Under Secretary of Defense for Policy and  Vice President and Director of Studies at CNAS [Michael, Paul, January 12 2021, “AI and International Stability: Risks and Confidence- Building Measures”, , Acc 6/18/22. M.A.]

Finally, autonomous systems raise novel challenges of signaling in contested areas because of ambiguity about whether their behavior was intended by human commanders. Even if the system performs as intended, adversaries may not know whether an autonomous system’s behavior was consistent with human intent because of the aforementioned command-and-control issues. This can create ambiguity in a crisis situation about how to interpret an autonomous system’s behavior. For example, if an autonomous system fired on a country’s forces, should that be interpreted as an intentional signal by the commanding nation’s political leaders, or an accident? This, again, is not a novel problem; a similar challenge exists with human-commanded military forces. Nations may not know whether the actions of an adversary’s deployed forces are fully in line with their political leadership’s guidance. Autonomous systems could complicate this dynamic due to uncertainty about whether the actions of an autonomous system are consistent with any human’s intended action.

2022 MNDI Negative
Neg › --Extend – Human Control Fails

Human control fails – our soldiers are not adequately trained for human/machine teams

Goldfarb and Lindsay, 2022 – Chair in AI and health care at the Univ of Toronto, Prof of Cybersecurity at the Georgia Tech [Avi, Jon, 2/25/22, , “Prediction and Judgement: Why Artificial Intelligence Increases the Importance of Humans in War” 6/18/22, LND]

A fundamental organizational challenge is to recruit, train, and retain the human talent required for human-machine teaming. We anticipate that AI systems will increase the influence of junior personnel, giving more leverage to their judgment and decisions. Yet, we also expect that the junior officers, noncommissioned officers, civilian employees, and government contractors who maintain and operate AI systems will struggle to understand the consequences of their actions in complex political situations. Gen. Charles Krulak highlights the role of “the strategic corporal” on twenty-first-century battlefields. Krulak argues that operational complexity makes tactical actions more strategically consequential, for better or worse, which places a premium on the character and leadership ability of junior personnel. AI will further increase the burden of judgment on them. Forward personnel will have to see the predictions from AI systems, assess whether the data that created the predictions are reliable, and make value judgments about how and why automated systems can advance the mission. Furthermore, AI systems will require constant reconfiguration and repair as the context of human-machine teaming changes during actual operations. Military personnel have long engaged in field-expedient, bottom-up innovation. We expect personnel will likewise hack AI systems to improve mission performance, as they understand it, even as unauthorized modifications put them into conflict with system configuration managers elsewhere in the bureaucracy. It is important to emphasize the human capital requirements of combining a sophisticated understanding of the politico-military situation with the technical savvy to engineer AI in the field. The strategic corporal in the AI era must be not only a Clausewitzian genius but also a talented hacker. This may not be a realistic requirement. The importance of human-machine teaming is increasingly appreciated in organizations that implement AI systems. Amid all the hype about AI and war, plenty of thoughtful work seeks to discern the relative advantages of humans and machines and to devise methods of pairing them together in order to improve decision-making. As the U.S. Department of Defense AI strategy states, “The women and men in the U.S. armed forces remain our enduring source of strength; we will use AI-enabled information, tools, and systems to empower, not replace, those who serve.” Yet, the strategy's stated goal of “creating a common foundation of shared data, reusable tools, frameworks and standards, and cloud and edge services” is more of a description of the magnitude of the problem than a blueprint for a solution. As AI creates potential for large-scale efficiency improvements, it also creates potential for large-scale collective action problems. New military staff specialties are sure to emerge to manage data and judgment resources, creating new institutional equities and integration challenges. Perhaps even more challenging is the problem of nurturing trust among all the engineers, administrators, analysts, operators, and lawyers involved in designing, using, and repairing AI systems. As cheap prediction makes human judgment more vital in a wide variety of tasks, and as more judgment is needed to coordinate human-machine teaming, we anticipate that military bureaucracies will face complicated command decisions about why, and how, to conjoin humans and machines. Commercial firms that embrace AI often adjust their boundaries and business models by contracting out tasks involving data, prediction, and action (e.g., manufacturing, transportation, advertising, and service provision) while developing in-house judgment capacities that are too difficult to outsource. Military organizations, likewise, may find it advantageous to share specialized resources (sensors, shooters, intelligence products, and logistics) across a decentralized network of units, even as they struggle to make sense of it all. AI is thus part of a broader historical trend that has been described with terms like “networkcentric warfare,” “joint force operations,” “integrated multi-domain operations,” and “interagency cross-functional teams.” The whole is more than the sum of its parts, but each part must exercise excellent judgment in how it leverages shared assets. Historical experience suggests that military interoperability and shared sensemaking are difficult, but not necessarily impossible, to achieve. We thus expect military and political judgment will become even more difficult, diffused, and geographically distributed. Indeed, the ongoing involvement of the “strategic corporal” in conversations about politico-military ends could end up politicizing the military. In the United States, as Risa Brooks argues, the normative separation of political ends from military means has some paradoxically adverse consequences: it enables service parochialism, undermines civilian oversight, and degrades strategic deliberation. Greater reliance on AI could exacerbate all these problems, precisely because AI is a force multiplier that requires military personnel to exercise greater judgment. Brooks's argument implies that an AI-intensive defense bureaucracy could become both more powerful and more politically savvy. If machines perform the bulk of data gathering, prediction, and tactical warfighting, then the judgments of human engineers, managers, and operators will be highly consequential, even as ethical questions of accountability become harder to answer. Some military personnel may be unable to perform at such a high level of excellence, as attested by the many scandals during the wars in Iraq and Afghanistan (from targeting errors to prisoner abuse). Increasing reliance on AI will magnify the importance of leadership throughout the chain of command, from civilian elites to enlisted service members.