Just after removing the "no military use" clause last week, artificial intelligence (AI) "leader" OpenAI announced that it is working with the Pentagon on multiple projects including cybersecurity tools, which is contrary to the startup's earlier ban on providing artificial intelligence to the military.
Anna Makanju, OpenAI's vice president of global affairs, said in an interview at the World Economic Forum in Davos on Tuesday that the company is working with the U.S. Department of Defense to develop open source cybersecurity software tools and has held preliminary talks with the U.S. government on ways to help prevent veteran suicide.
Last week, the company quietly revised its "user policy" and removed the restriction that "prohibits the use of the company's AI for military and war purposes," sparking public opinion. Previous terms explicitly prohibited the use of company models in "high-risk activities causing physical harm," including weapons development, military and warfare.
The word "military" very clearly excludes the US Department of Defense, the US military, the Central Intelligence Agency (CIA), etc. from the category of users of OpenAI.
Makanju said the decision was part of a larger policy update to accommodate new uses for ChatGPT and its other tools.
“Because we had basically a blanket ban on military use before, many people thought that would ban a lot of these use cases that people thought were very consistent with what we wanted to see in the world,” she said. “But OpenAI insisted on prohibiting the use of its technology to develop weapons, destroy property, or harm people.”
This decision is actually not surprising, after all, defense purposes are more profitable. Microsoft, the largest investor in OpenAI, has also provided several software contracts to the U.S. military and other government agencies.
OpenAI also said it is accelerating its work on election security, devoting resources to ensuring its generative AI tools are not used to spread political disinformation.