NATO AI Strategy

Posted on Updated on

Over the past 18 months there has been wide spread push buy many countries and geographic regions, to examine how the creation and use of Artificial Intelligence (AI) can be regulated. I’ve written many blog posts about these. But it isn’t just government or political alliances that are doing this, other types of organisations are also doing so.

NATO, the political and (mainly) military alliance, has also joined the club. They have release a summary version of their AI Strategy. This might seem a little strange for this type of organisation to do something like this. But if you look a little closer NATA also says they work together in other areas such as Standardisation Agreements, Crisis Management, Disarmament, Energy Security, Clime/Environment Change, Gender and Human Security, Science and Technology.

In October/November 2021, NATO formally adopted their Artificial Intelligence (AI) Strategy (for defence). Their AI Strategy outlines how AI can be applied to defence and security in a protected and ethical way (interesting wording). Their aim is to position NATO as a leader of AI adoption, and it provides a common policy basis to support the adoption of AI System sin order to achieve the Alliances three core tasks of Collective Defence, Crisis Management and Cooperative Security. An important element of the AI Strategy is to ensure inter-operability and standardisation. This is a little bit more interesting and perhaps has a lessor focus on ethical use.

NATO’s AI Strategy contains the following principles of Responsible use of AI (in defence):

  • Lawfulness: AI applications will be developed and used in accordance with national and international law, including international humanitarian law and human rights law, as applicable.
  • Responsibility and Accountability: AI applications will be developed and used with appropriate levels of judgment and care; clear human responsibility shall apply in order to ensure accountability.
  • Explainability and Traceability: AI applications will be appropriately understandable and transparent, including through the use of review methodologies, sources, and procedures. This includes verification, assessment and validation mechanisms at either a NATO and/or national level.
  • Reliability: AI applications will have explicit, well-defined use cases. The safety, security, and robustness of such capabilities will be subject to testing and assurance within those use cases across their entire life cycle, including through established NATO and/or national certification procedures.
  • Governability: AI applications will be developed and used according to their intended functions and will allow for: appropriate human-machine interaction; the ability to detect and avoid unintended consequences; and the ability to take steps, such as disengagement or deactivation of systems, when such systems demonstrate unintended behaviour.
  • Bias Mitigation: Proactive steps will be taken to minimise any unintended bias in the development and use of AI applications and in data sets.

By acting collectively members of NATO will ensure a continued focus on interoperability and the development of common standards.

Some points of interest:

  • Bias Mitigation efforts will be adopted with the aim of minimising discrimination against traits such as gender, ethnicity or personal attributes. However, the strategy does not say how bias will be tackled – which requires structural changes which go well beyond the use of appropriate training data.
  • The strategy also recognises that in due course AI technologies are likely to become widely available, and may be put to malicious uses by both state and non-state actors. NATO’s strategy states that the alliance will aim to identify and safeguard against the threats from malicious use of AI, although again no detail is given on how this will be done.
  • Running through the strategy is the idea of interoperability – the desire for different systems to be able to work with each other across NATO’s different forces and nations without any restrictions.
  • What about Autonomous weapon systems?  Some members do not support a ban on this technology.
  • Has similar wording to the principles adopted by the US Department of Defense for the ethical use of AI.
  • Wants to make defence and security a more attractive to private sector and academic AI developers/researchers.
  • NATO principles have no coherent means of implementation or enforcement.