Pentagon to Look Into Generative AI for National Security

The US Department of Defense, DoD, announced the creation of a task force to study the possibilities of using generative AI. The decision comes amid huge public attention to generative AI programs such as ChatGPT and Google Bard.

The Pentagon plans to explore how this technology can be applied to national security, intelligence services, industry, business, and health.

Possibilities of AI technologies

Deputy Secretary of Defense Dr. Kathleen H. Hicks will play a key role in the analysis and integration of generative AI tools. Hicks says the creation of such a group speaks to the DoD’s commitment to innovation, emphasizing that the future of defense, along with the introduction of advanced technologies, requires foresight and responsibility.

The Task Force Lima will be led by the Chief Digital and Artificial Intelligence Office (CDAO). Specialists will evaluate, synchronize, and use the generative capabilities of artificial intelligence.

“There’s a little bit of fearmongering” around AI, the Pentagon’s Chief Digital & AI Officer tells me, arguing it’s neither panacea nor Pandora’s box.

“It’s not like we’ve gotten to the point where… there’s some monolithic intelligence – which is the way it’s being pitched.” pic.twitter.com/fVwIOgH5GC

— Christiane Amanpour (@amanpour) August 7, 2023

According to US Navy Captain M. Xavier Lugo, Task Force Lima mission commander and member of the CDAO’s Algorithmic Warfare Directorate, the use of artificial intelligence in defense is also related to strengthening national security. He noted that the DoD recognizes the potential of generative AI to significantly improve intelligence, operational planning, and administrative and business processes. At the same time, Lugo called the responsible adoption of technology the key to effective management of associated risks.

Craig Martell, the DoD Chief Digital and AI Officer, also emphasized the importance of the responsible approach. The DoD should determine appropriate safeguards and mitigate the national security risks that can result from issues such as poorly managed training data, he said. He also noted the need to consider the extent to which the US adversaries would use the same technology.

Uncertainty related to AI

Artificial intelligence, which belongs to the category of generating AI algorithms, creates new outcomes from training data, diverging significantly from simpler machine learning methods. AI works with structured data for reliable statistical outcomes. Accessible generative tools of artificial intelligence, such large language models including ChatGPT, create text nearly indistinguishable from human speech. However, because these models learn from vast datasets, including the internet, they can occasionally generate “hallucinations” or inaccuracies. Thus, the Pentagon has previously expressed caution in applying generative AI.

Quantifying the impact of unreliability on risk and finding the right balance between pre-trained data and “hallucinations” remains uncertain. A task force seeks industry collaboration to align products with DoD standards, emphasizing risk reduction and user-friendliness.

About CDAO

Established in June 2022, CDAO is spearheading the integration and enhancement of artificial intelligence capabilities within the US DoD. Its pivotal role revolves around expediting the Department’s assimilation of data, analytics, and AI in order to fortify the digital framework and policies.