2 ms·

> "Signatories who integrate generative AI into their services like Bingchat for Microsoft, Bard for Google (GOOGL.O) should build in necessary safeguards that

> "Signatories who integrate generative AI into their services like Bingchat for Microsoft, Bard for Google (GOOGL.O) should build in necessary safeguards that these services cannot be used by malicious actors to generate disinformation," Jourova told a press conference. Basically you can use AI generated content to promote what the government likes. However, if you go against what the government likes, you can’t use AI.

ben_w 3y ago
I'm sure some people have such cynical reasons for suggesting that as a "solution"; one other possibility is that it's magical thinking — it wouldn't be the first time I've heard someone say something like "just make it tell the truth" as if that's a thing we can always know and recognise ourselves as humans. And OpenAI has warned about this specific failure mode since at least 2018: https://openai.com/research/preparing-for-malicious-uses-of-ai https://openai.com/research/preparing-for-malicious-uses-of-...
octopoc 3y ago
I predict that as time moves on, AI will be responsible for generating a larger and larger percentage of politically correct content, until there comes a day when, for all practical purposes, 100% of all politically correct content is AI generated and 100% of politically incorrect content is created organically. At that point there will be two worlds, one that feels genuine and is politically incorrect and the other that feels like a fairy tale and is oriented around farming consumers.
briandear 3y ago
And on top of that, a human can create "bad" content. So the means by which it's created is a bit irrelevant. And if we want to talk about "disinformation" -- how about all of the disinformation created by governments themselves to suit their specific agenda? And beyond the actual government -- the disinformation from political candidates is an art form.

ben_w 3y ago

I'm sure some people have such cynical reasons for suggesting that as a "solution"; one other possibility is that it's magical thinking — it wouldn't be the first time I've heard someone say something like "just make it tell the truth" as if that's a thing we can always know and recognise ourselves as humans. And OpenAI has warned about this specific failure mode since at least 2018: https://openai.com/research/preparing-for-malicious-uses-of-ai https://openai.com/research/preparing-for-malicious-uses-of-...

octopoc 3y ago

I predict that as time moves on, AI will be responsible for generating a larger and larger percentage of politically correct content, until there comes a day when, for all practical purposes, 100% of all politically correct content is AI generated and 100% of politically incorrect content is created organically. At that point there will be two worlds, one that feels genuine and is politically incorrect and the other that feels like a fairy tale and is oriented around farming consumers.

briandear 3y ago

And on top of that, a human can create "bad" content. So the means by which it's created is a bit irrelevant. And if we want to talk about "disinformation" -- how about all of the disinformation created by governments themselves to suit their specific agenda? And beyond the actual government -- the disinformation from political candidates is an art form.