The US Government Secretly Tried to Use AI to Spy on Americans and the Company Said No

Here's a story that should make you uncomfortable: according to MIT Technology Review, the US government wanted to use Anthropic's Claude AI to analyze massive amounts of data collected from everyday Americans.

Let that sink in. The government approached one of the most powerful AI companies in the world and basically said, 'Hey, can we use your super-intelligent chatbot to sift through data we've been collecting on our own citizens?'

This revelation comes at a weird time, because Anthropic's CEO Dario Amodei just went on camera declaring 'We are patriots' in an interview that's gotten over 53,000 views on YouTube. The timing of that statement hits different when you learn the government was knocking on their door asking for surveillance help.

To be clear, there's no indication Anthropic agreed to the request. But the fact that the ask was made at all raises huge questions. What kind of data are we talking about? How was it collected? And if Anthropic said no, did the government just move on to the next AI company?

This is the kind of stuff that privacy advocates have been screaming about since AI got powerful enough to process billions of data points in seconds. A human analyst can only read so many documents. An AI can read all of them.

As reported by MIT Technology Review.

Source: MIT Technology Review