Facebook whistleblower Frances Haugen issues chilling warning about AI and says it could soon have ‘civilisation-altering impacts’

A former Facebook employee has warned that advances in artificial intelligence could have “civilization-changing effects” and rapidly increase the amount of dangerous misinformation spreading online.

Whistleblower Frances Haugen said that as artificial intelligence grows in scale and economies rely more on software running in data centres, the world will begin to see the “darkening age” creeping in.

The former engineer and product manager — who left Facebook in 2021 after thousands of documents were leaked showing toxic content was intentionally spread through the platform — said that without stronger regulation there would be “a repeat of what we saw with social media” on a much larger scale. size.

“When we start getting into scalable systems that run on data centres, a very small number of people can have civilization-altering levels of power,” Ms Haugen told the National Press Club on Tuesday.

“At Facebook, there are very few people who really understand how these algorithms work and yet they influence what everyone sees in the news.

“When we see the kind of scale that can impact, that can have really significant consequences.”

As artificial intelligence grows in scale and economies become more reliant on software running in data centres, the world will begin to see the “Dark Age” creeping towards us, said whistleblower Frances Haugen.

Millions of people around the world use AI software every day.

ChatGPT has over 100 million users worldwide. It achieved 1.6 billion visits in June.

Recent surveys suggest that nearly one in four Australians use AI software in their workplace, and up to 70% of 14- to 17-year-olds have used AI software at least once.

Ms Haugen – who built a successful tech career in California’s Silicon Valley before testifying that Facebook promoted hate speech and pushed eating disorder content to teens on Instagram – said newer algorithms promoted “the most divisive and polarizing content” for profit.

Content that keeps you around has been the best content. Then they said that if you get a reshare, like, or comment, that’s good content.

“Now, our most divisive, polarizing, and worst content gets the most distribution.

“Unless we pass laws that say, ‘Hey, we’re not going to repeat the mistakes of social media, we’ve got to be transparent…we’ve got to protect whistleblowers,’ we’re going to see a repeat of what we saw with Facebook.”

“Anytime there’s a profit motive and there’s no feedback loop to correct those lies, we’re going to see those gaps get wider and wider.”

The number of ChatGPT users rose from 100 million in January to 180.5 million by August. Photograph: Marco Bertorello/AFP

In the wake of calls for stronger regulation of online hate speech and misinformation, protections for whistleblowers and the media are clearly failing in Australia, according to one of the most prominent free expression advocates.

Award-winning journalist Peter Gresty said Australia needed a mechanism that took the role of journalists and whistleblowers into account.

He also said the government and Australian Defense Force needed to be “more transparent” about the work they had done, especially amid the deepening crisis in the Middle East.

Professor Gristy told the National Press Club: “I think that at the moment, generally speaking, as a country and particularly for our security and defense agencies, we are too obsessed with secrecy, with keeping these matters closed.”

“I think this causes a great deal of problems for everyone.”