OpenAImaker of ChatGPT and one of the world's leading artificial intelligence companies, today said it has partnered with Anduril, a defense startup that makes missiles, drones and software for the US military. It marks the latest in a series of similar announcements recently made by major tech companies in Silicon Valley, which has forged increasingly closer ties with the defense industry.
“OpenAI is building AI that benefits as many people as possible, and supports US-led efforts to ensure the technology upholds democratic values,” Sam Altman, CEO of OpenAI, said in a statement on Wednesday.
OpenAI's AI models will be used to improve air defense systems, Brian Schimpf, co-founder and CEO of Anduril, said in the statement. “Together we are committed to developing responsible solutions that enable military and intelligence agencies to make faster and more accurate decisions in high-pressure situations,” he said.
OpenAI's technology will be used to “assess drone threats faster and more accurately, giving operators the information they need to make better decisions while staying out of harm's way,” said a former OpenAI employee who joined the company earlier this year left and spoke on the condition of anonymity to protect their professional relationships.
Earlier this year, OpenAI changed its policy regarding the use of its AI for military applications. A source who worked at the company at the time says some employees were unhappy with the change, but there were no open protests. According to reporting from The Intercept, the US military is already using OpenAI technology.
Anduril is developing an advanced air defense system with a swarm of small, autonomous aircraft that work together during missions. These aircraft are controlled via an interface powered by a large language model, which interprets and translates natural language commands into instructions that both human pilots and the drones can understand and execute. So far Anduril has used open source language models for testing purposes.
Anduril is currently not known to use advanced AI to control its autonomous systems or allow them to make their own decisions. Such a move would be riskier, especially given the unpredictability of current models.
A few years ago, many AI researchers in Silicon Valley were strongly opposed to working with the military. In 2018, thousands of Google employees staged protests against the company that provided AI to the US Department of Defense through what was then known within the Pentagon as Project Maven. Google later withdrew from the project.
