Free beta access. 5 free AI tokens each week. Optional top-ups available. See what is included
Opening Word document…

AI increases threats to cyber security because it increases the dependence on data, while creating new targets of data attacks.

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, https://doi.org/10.1162/isec_a_00425, “Prediction and Judgement: Why Artificial Intelligence Increases the Importance of Humans in War” 6/18/22, LND]

In short, there may be more data in modern war, but data management has also become more challenging. Although U.S. weapons may be fast and precise, U.S. wars in recent decades have been protracted and ambiguous. We argue that AI will most likely deepen rather than reverse these trends. Indeed, automation is both a response to and a contributing cause of the increasing complexity of military information practice.

Just as commanders are already preoccupied with C2 architecture, officers in AI-enabled militaries will seek to gain access to large amounts of data that are relevant to specific tasks in order to train and maintain AI systems. Units will have to make decisions about whether they should collect their own data organically or acquire shared data from other units, government agencies, or coalition partners. We expect many relevant databases to be classified and compartmented given the sensitivity of collection techniques or the content itself, which will complicate sharing. Units might also choose to leverage public data sources or purchase proprietary commercial data, both of which are problematic because nongovernmental actors may affect the quality of and access to data. As militaries tackle new problems, or new operational opportunities emerge, data requirements will change, and officers will have to routinely find and integrate new data sources. AI strategy will require militaries to establish data policies, and thus negotiating access to data will be an ongoing managerial—and human—challenge. We contend that militaries will face not only data access but also data relevancy challenges. Heterogenous data-generating processes allow biases and anomalies to creep into databases. Although metadata may help to organize information processing, they are also vulnerable to data friction that only humans can fix. Cleaning and curating data sources will therefore be as important as acquiring them in the first place. To the challenges of producing or procuring data must be added the challenges of protecting data. Just as supply chains become attractive targets in mechanized warfare, data supplies will also become contested. Overall, we expect the rise of AI to exacerbate the already formidable challenges of cybersecurity. Cybersecurity professionals aim to maintain the confidentiality, integrity, and availability of an organization's data. Two of these goals—integrity and availability—capture the AI requirements of unbiased and accessible data, as discussed above. The goal of confidentiality is also important insofar as data provide AI adopters with a competitive advantage. In commerce, AI companies often try to own (rather than buy) the key data that enable their machines to learn. The military equivalent of this is classified information, which is hidden from the enemy to produce a decision advantage. Military organizations will have strong incentives to protect the classified data that military AI systems use to learn. For the same reasons, adversaries will have incentives to steal, manipulate, and deny access to AI learning data. To date, most discussions of AI and cybersecurity have focused on a substitution theory of cybersecurity, that is, using AI systems to attack and defend networks. But we argue that a complementary theory of cybersecurity is just as if not more important. AI will require the entire military enterprise to invest more effort into protecting and exploiting data. If AI systems are trained with classified information, then adversaries will conduct more espionage. If AI enhances intelligence, then adversaries will invest in more counterintelligence. If AI provides commanders with better information, then adversaries will produce more disinformation. Inevitably, different parts of the bureaucracy will tussle among themselves and with coalition partners and nongovernmental actors to access and curate a huge amount of heterogeneous and often classified data. Organizations will also struggle with cyber and intelligence adversaries to maintain control of their own data while also conducting their own campaigns to collect or manipulate the enemy's data. To appreciate the strategic implications of AI, therefore, it is helpful to understand cyber conflict, most of which to date resembles espionage and covert action more than traditional military warfare. In fact, chronic and ambiguous intelligence contests are more common than fast and decisive cyberwar. Military reliance on AI becomes yet another factor abetting the rise of cyber conflict in global affairs, and the (ambiguous, confusing, interminable, gray zone) dynamics of cyber conflict are likely to have a strong influence on the dynamics of AI conflict.

The Impact - Cyber attacks escalate to nuclear war due to the fog of war – simulations prove

Schneider, 2022 - Fellow at the Hoover Institution at Stanford University [Jacquelyn; March 7; Affairs, “The Biggest Cyber Risk in Ukraine? How Russian Hacking Could Threaten Nuclear Stability,” from Cyber Starter Packet]

Why the apparent restraint? It is almost impossible to know exactly why (or if) the Russians have indeed held back. Perhaps cyber-operations have been attempted and failed; perhaps Russian President Vladimir Putin has held his cyber-capabilities in reserve, saving them for later. Or maybe cyber-operations have taken place, but their effect—which is often virtual and not clearly attributed—will take longer to materialize. What is known is that the conflict is far from over, and the next question becomes whether cyber-operations could play a larger role as the war turns more violent. It is likely that the next stage of conflict will more than ever be defined by planes, tanks, artillery, and soldiers. It seems unlikely, given the amount of indiscriminate damage currently being inflicted by Russia, that cyber-operations will escalate the violence of the campaign within Ukraine. That said, could cyber-operations lead to horizontal escalation, drawing NATO into the fight, for example? Or, given that the United States and Russia are the world’s largest nuclear powers, could cyber-operations escalate to the worst possible outcome—nuclear war? Recent wargaming research suggests that cyber-exploits into nuclear command and control may be enticing for states looking to neutralize a nuclear escalation threat in the midst of a conventional war, and that actors may underestimate the danger of these exploits and vulnerabilities to nuclear stability. GETTING PULLED IN One way cyber-operations could lead to escalation is by pulling the United States or NATO into the conflict. Mark Warner, the Democratic senator from Virginia, warned in late February that potential Russian cyberattacks on critical infrastructure in Ukraine could have accidental spillover effects on NATO countries—for instance if a Russian cyberattack on Ukrainian energy infrastructure caused an outage in a NATO neighbor like Poland. This could inadvertently trip Article 5 of NATO’s founding treaty, which states that an armed attack against one member state will be considered an attack against them all. This would be uncharted waters for NATO, which only recently publicly stated that cyberattacks might invoke Article 5 and is still ambiguous about what types of cyberattack—which range from virtual outages to data manipulations to physical damage (in extremely rare circumstances)—might be serious enough for NATO to respond with conventional retaliation. The Biden administration has warned that the United States would respond to cyberattacks on U.S. critical infrastructure, such as the country’s electrical grid or water supply (although officials stopped short of saying how the United States would respond). So far, the United States has answered previous cyberattacks with either sanctions, law enforcement actions, or the confiscation of cryptoassets. None of these options seem likely to deter Putin at this point, and so the Biden administration may find itself in an unprecedented position of having few credible options to threaten Russia. It is certainly possible that Putin, facing a conventional war that he thinks he might lose, could attack critical infrastructure in the United States or other NATO countries in the hope that their citizens will push their governments to abandon Ukraine. The financial sector, in particular, would seem to be a logical target for Russian cyberattacks, given the damage that Western economic actions have already done to the Russian economy. It is difficult to create widespread and long-lasting effects with cyberattacks, however, and the financial sector is the best equipped and most advanced cyber-defender in the world. Plus, research I’ve conducted with Sarah Kreps, director of the Cornell Tech Policy Lab, finds that the American public views cyberattacks as qualitatively different from conventional means of warfare—more akin to economic sanctions than bombs. Thus, cyberattacks are unlikely to provoke the kind of retaliation or emotional response that would pull the United States or its NATO allies into a war with Russia. What’s more, the United States can probably withstand the short-term damage to critical infrastructure that a Russian cyberattack might create, and such attacks might actually increase resolve to support Ukraine. This means a deliberate choice by Russia to use cyberattacks against the United States or NATO to “escalate to dominate”—deliberately ratcheting up the pressure to force Washington to back off—would likely fail. A more troubling scenario involves accidental escalation from cyber-operations—that is, when critical infrastructure is unintentionally damaged by a cyberattack or when a cyberattack is misattributed to Russia (or the United States). This is especially dangerous for civilian infrastructure that also serves military or security purposes—for example, harming a refugee train by using a cyberattack targeting railroads also used to move troops and supplies to the front. Plus, a jumble of actors has jumped into this space, from criminal syndicates to cyber-militias to hacker collectives such as Anonymous. That increases the chances that one of these players will target civilian infrastructure, and misattribution to either Russia or the United States could needlessly trigger retaliation. WHEN CYBER GOES NUCLEAR By far the most dangerous form of escalation is the possibility that a cyber-operation increases the likelihood of nuclear war. How likely is such a scenario? No one may know if Russia has a cyberweapon that can target nuclear weapons (or, for that matter, whether the United States does), but there are theories and some data about how the cyber-realm might affect nuclear stability. American policymakers have generally recognized that attempting to interfere with nuclear command, control, and communications could lead to dangerous incentives for states to launch nuclear weapons preemptively. Threats to nuclear command and control, for example, could leave states so fearful about their second-strike capability (the ability to launch a nuclear weapon in retaliation against an attacker) that in the midst of a conflict they would feel compelled to use nuclear weapons preemptively. Some scholars have warned that attacks against nuclear command-and-control systems could make it impossible to control nuclear war and keep it limited, leading to inadvertent nuclear Armageddon. Despite these fears about the dangers of attacking nuclear command and control, there was never an agreement between the United States and the Soviet Union (and subsequently Russia) to not attack each other’s nuclear command, control, and communications.

Sign in to recut, update, or save this card.