The EU is not implementing military AI policies now – they are focused on regulating civilian technology
Stanley-Lockman, 2021 - Center for Security and Emerging Technology [Zoe CSET Issue Brief August “Responsible and Ethical Military AI Allies and Allied Perspectives” https://cset.georgetown.edu/publication/responsible-and-ethical-military-ai/ Acc. 4/21/22 TA]
While the EU has adopted a bullish approach to trustworthy AI in the civilian realm, European institutions have been slower to define the implications for safe and ethical AI beyond the tip of the spear. Key civilian policies and regulations, like the General Data Protection Regulation and more recent legislation instituting the European approach to “trustworthy” AI, have clear carveouts for public safety, security, and defense. Still, the EU approach to civilian AI policy is relevant to transatlantic defense because the dual-use, general-purpose nature of AI means that military adoption of AI will depend on the ethical frameworks that dominate civilian development, regardless of carveouts. More directly, some European countries also choose to apply EU legislation like the General Data Protection Regulation to their own defense sectors, even though they are not required to do so.167 With this overlap in mind, Appendix VIII overviews the applicability of the EU trustworthy AI principles for the defense realm. In addition to examples such as Airbus’ application of the ALTAI methodology to FCAS, it is notable that several European defense efforts mention the European Commission-supported guidelines for trustworthy AI as a positive step toward ensuring military uses of AI adhere to ethical standards. For example, in their co-authored food-for-thought paper on AI in defense, Finland, Estonia, France, Germany, and the Netherlands made explicit reference to the Trustworthy AI Principles, recognizing that the EU could leverage its normative power because of the centrality of ethical standards in AI for defense.168 The focus on safety and security in EU AI policy also promotes “convergence between the AI community and the security community” to enhance robustness.169 In sum, the emphasis on safety, security, and risk in EU AI policy is not only a natural overlap, but also one that European defense stakeholders are seeking out. However, it remains to be seen which EU body will take control of the responsible and ethical military AI agenda. There are various actors within the EU institutions that are largely beyond the scope of this paper.170 Instead, there are only inklings of how the EU will approach responsible and ethical military AI at present. In the future, this topic could also feature in EU-U.S. security and defense dialogues. For now, it is the European Parliament that plays the most visible role advancing ethics in European military R&D funding. This was seen in mid-2018, in its attempt to ban all military AI research using EU funds because of concerns about LAWS. The agreed upon final version explicitly prohibits funding for LAWS at the European level—a deal-breaker without which the Parliament would never have agreed to allow for any defense funding.171 But the final result was narrower because the EU does not have jurisdiction over its member states’ armaments development unless they use EU funds. As such, while important, especially for dualuse and open-source systems, its jurisdiction on mandating ethical reviews is still limited and likely to not affect the majority of national, bilateral, or minilateral capability development programs. More recently, the Parliament-issued Guidelines for military and non-military use of Artificial Intelligence in January 2021 could indicate a stronger ethical bent than seen in the other institutions.172