Google is joining the push to weaponize AI

The next live-action sequel in the Terminator franchise could soon play out in real time.

Google updated its ethical guidelines around AI this week, removing a company-wide pledge to avoid using the technology to develop potentially harmful products like weapons or surveillance.

• This pledge dates back to Google’s first AI principles published in 2018, which came after employees protested a contract with the Pentagon applying algorithms to analyze drone footage.
• In a blog post, top Google officials said the change stems from the belief that “companies, governments, and organizations sharing [core democratic] values should work together to create AI that protects people, promotes global growth, and supports national security.”

AI is already being adapted for the battlefield

Google’s previous restrictions on using AI for national security applications had made it an outlier among major competitors.

• OpenAI is working with military manufacturer Anduril to develop technology for the Pentagon.
• Anthropic has partnered with defense contractor Palantir to help US intelligence and defense agencies access versions of its chatbot Claude.
• Tech giants Microsoft and Amazon have long partnered with the Pentagon, including on AI-related projects.

🤖 Use cases abound: As of last spring, the US military’s AI-controlled fighter jets were “roughly even” in skill level compared to experienced human pilots not named Pete “Maverick” Mitchell, according to Air Force Secretary Frank Kendall. Additionally, a wide range of low-cost drones can use AI image recognition to identify and surveil targets – or blow them up.

Outside of aircraft, the US military and its contractors have spent billions developing experimental submarines, tanks, ships, and other technology that use AI to pilot themselves and shoot – including robot dogs armed with rifles.