Military training AI to be its “Dr. Watson.” What could go wrong?

The US military is increasingly using AI applications. Experts are worried, of course.

The 15th Marine Expeditionary Unit, which spent most of 2024 aboard ships throughout the Pacific, has tested a leading AI tool funded by the Pentagon.

The new system was used to gather and interpret open-source intelligence. It scouted non-classified articles, reports, images, and videos, and later summarized foreign news sources, and wrote daily and weekly intelligence reports, the MIT Technology Review reported.

The tools used by the Marines were built by Wannevar Labs, and are part of a wider Pentagon project to spend $100 million in the next two years on AI pilots.

As per Vannevar Labs’ chief technology officer Scott Philips, the real focus of the company is to “help the US make good decisions.”

The Pentagon is also working with Microsoft and Palantir on AI models that would help with data analysis.

Rapid AI tool implementation is worrisome. Experts quoted by the MIT Technology Review worry that overreliance on AI could lead to unnecessary escalation of situations and misinformed decisions, among other negative consequences.

You can read the full article here.