Google DeepMind staff quit over Pentagon AI deal

Google DeepMind staff have quit over the company’s military AI deals with the Pentagon. Some fear the technology could be used in warfare without proper checks. Their exits raise a bigger question, can AI companies keep control over how their tools are used?

Why Google DeepMind staff are leaving

As reported by NDTV, Sofia Liguori, a research engineer based in London, resigned from Google DeepMind after three years at the company. She said its growing work with the US and Israeli militaries had made her lose faith in its stated values. She also said she felt ‘ashamed to work‘ at Google as it pushed further into military AI.

AI researcher Andreas Kirsch also left earlier in October. He questioned whether Google had enough checks to make sure its AI was used safely in secret military settings. He said the company seemed to have little power to stop the Pentagon from using its tools in ways it could not fully see. The workers are concerned about the business deals signed this year. These allow AI systems to be used across classified US military networks.

Why military AI worries the former staff

AI tools can sort large amounts of data and help people make decisions faster. But they can also bring risks if a system makes a mistake or if people trust its output too much. Staff raised concerns about weak human oversight and the lack of clear limits on the Pentagon’s use of the technology. They fear that companies could use AI in high stakes decisions without giving people enough time to check its results.

One example is Maven Smart Systems, a Pentagon platform that uses AI to help organise intelligence and identify possible targets. It was developed by defence firm Palantir and has brought together tools from several AI companies. Critics say faster decisions are not always safer decisions, especially when lives may be at stake. The Pentagon has argued that AI is important to modern military work.

Google defends its DeepMind contracts

Google says it reviews public sector deals under its policies and AI principles. A company spokesperson said these contracts must follow its rules for acceptable use. Google also says staff have safe ways to raise concerns and that it does not allow retaliation at work. Google’s AI principles include human oversight and respect for international law and human rights. However, some former workers say they are not sure these promises provide enough control when companies use AI in classified settings.

The debate is not new. In 2018, employee protests led Google to step away from Project Maven, a military project that used AI to study drone footage. This time, the company has agreed to let the Pentagon use its AI in classified work, under terms that permit “any lawful government purpose” while setting some limits on mass surveillance and autonomous weapons. Google is among several technology firms that have signed deals to supply AI tools for classified military work.

Author’s Opinion

AI can help armies process data and plan tasks, but speed alone cannot make warfare safer. Companies must explain how people may use their tools and what safeguards they apply. Workers should be able to raise concerns when they believe human lives could be at risk.

ALSO READ: Apple iPhone Duo buyers face a major app problem at launch.