OpenAI says its new robot-writer is too dangerous to be published


OpenAI, a nonprofit organization for AI, has come up with a text generator so effective at creating "deep news" that its creators have decided that the program too dangerous to make public.

OpenAI's writing will not be in your Facebook feed soon, but Robo's authors are already helping other companies write, making it harder than ever for regulators to curb false news.

OpenAI-ing Pandora's Box

In 2015, Sam Altman and Elon Musk expressed concern that the most powerful AI programs in the world were all developed behind closed doors. That's why they launched OpenAI, a non-profit organization, whose mission is to make a "safe" artificial intelligence public.

But the OpenAI program, called GPT2, is so good that it produces a writing virtually indistinguishable from true journalism, opening the door to more and more sophisticated false news.

"It's very clear that if this technology matures, it could be used for misinformation or propaganda purposes," said OpenAI's director of policy, Jack Clark, Review of MIT technology. "We try to get in advance."

The reporters are already writing

OpenAI's text generator will be locked until its creators understand what it can and can not do.

But, robo-writers are already roaming: Bloomberg News uses a robot-writing program called Cyborg in ⅓ of his articles, and The Washington Post, the Associated press, and The Guardian all produce "machine-assisted" writing.

GPT2 – or something like that – will eventually become public. The researchers hope they will have a way to control it. "We try to build the road while we travel …" said Clark The Guardian.

It sounds like something that a robot would write …

Source link