NATO Updates AI Strategy to Address Generative AI, Disinformation

NATO has revised its AI strategy to keep up with the changing technology landscape, adding provisions related to generative AI and AI-enabled disinformation.

First published in 2021, NATO’s AI strategy sets out how the defense organization will collaborate and cooperate on matters related to AI for transatlantic defense and security.

In updates to the strategy, NATO said it will “work to protect against the adversarial use of AI, including through increased strategic foresight and analysis.”

Newly included in the strategy are provisions for AI-generated disinformation.

Amid the rise of deepfakes targeting politicians, NATO added a section to its strategy to protect against the adversarial use of AI.

Examples of AI-generated disinformation that could impact the alliance’s efforts include a deepfake version of Ukrainian President Volodymyr Zelensky giving the order to surrender and ChatGPT maker OpenAI said in May that it uncovered at least five operations using its technologies to spread misinformation.

NATO’s strategy suggested that AI-enabled disinformation operations “might affect the outcome of elections, sow division and confusion across the alliance.”

The revised strategy also warns that AI technologies could be used to facilitate gender-based violence and to demoralize societies and militaries in times of conflict.

Related:NATO $1B Innovation Fund Makes First Investments in AI, Robotics

To protect against AI disinformation, the strategy stresses a need for “strategic foresight” which includes implementing proactive measures from “anticipatory governance to alternative scenario planning based on participatory and responsible approaches.”

“Under the new AI strategy, NATO will work to protect against the adversarial use of AI, including through increased strategic foresight and analysis,” the alliance said.

Ensuring Responsible AI in the Military 

NATO’s AI strategy is designed to help alliance members ensure they innovate responsibly while addressing security considerations and managing related risks.

Its outcomes now include an intent to increase the range of AI use cases suitable for members to use as well as revised steps to take during the integration of such technologies.

The strategy also includes new outcomes designed to improve NATO’s AI readiness, including an alliance-wide AI testing and evaluation program and increased standards and tools to review systems and best practices.

Alliance defense stakeholders will be required to ensure they’re responsibly adopting AI technologies.

NATO has long sought to ensure AI is used responsibly. Its chief, Jens Stoltenberg, previously suggested the alliance wants to develop shared standards with China to ensure responsible use.

Related:NATO Wants Deal with China on Responsible AI for Defense

To foster responsible technology uses among its members, NATO unveiled its own AI certification standard in February 2023, designed to ensure alliance suppliers and member states are deploying AI technologies in line with international law, as well as the organization’s values.

The newly revised strategy also suggests NATO and its alliances will need to be able to access specialized laboratories and sandboxes to test emerging AI capabilities. It also states that member states should look to conduct joint AI research projects alongside industry and academia to “help tap into collective expertise and resources. “

NATO’s strategy looks to address AI impacts on the military and civilian workforce, with the potential for upskilling programs referenced in the strategy.