Microsoft’s AI boss thinks it’s perfectly OK to steal content if it’s on the open web

5 months ago 81

Microsoft AI brag Mustafa Suleyman incorrectly believes that the infinitesimal you people thing connected the unfastened web, it becomes “freeware” that anyone tin freely transcript and use.

When CNBC’s Andrew Ross Sorkin asked him whether “AI companies person efficaciously stolen the world’s IP,” helium said:

I deliberation that with respect to contented that’s already connected the unfastened web, the societal declaration of that contented since the ‘90s has been that it is just use. Anyone tin transcript it, recreate with it, reproduce with it. That has been “freeware,” if you like, that’s been the understanding.

Microsoft is presently the people of multiple lawsuits alleging that it — and OpenAI — are stealing copyrighted online stories to bid generative AI models, truthful it whitethorn not astonishment you to perceive a Microsoft exec support it arsenic perfectly legal. I conscionable didn’t expect him to beryllium truthful precise publically and evidently wrong!

I americium not a lawyer, but adjacent I tin archer you that the infinitesimal you make a work, it’s automatically protected by copyright successful the US. You don’t adjacent request to use for it, and you surely don’t void your rights conscionable by publishing it connected the web. In fact, it’s so hard to waive your rights that lawyers had to travel up with special web licenses to help!

Fair use, meanwhile, is not granted by a “social contract” — it’s granted by a court. It’s a ineligible defence that allows some uses of copyrighted worldly erstwhile that tribunal weighs what you’re copying, why, however much, and whether it’ll harm the copyright owner.

That surely hasn’t kept galore AI companies from claiming that grooming connected copyrighted contented is “fair use,” but astir haven’t been arsenic brazen arsenic Suleyman erstwhile talking astir it.

Speaking of brazen, he’s got a prime punctuation astir the intent of humanity soon aft his “fair use” remark:

What are we, collectively, arsenic an organism of humans, different than a cognition and intelligence accumulation engine?

Suleyman does look to deliberation there’s thing to the robots.txt idea — that specifying which bots can’t scrape a peculiar website wrong a substance record mightiness support radical from taking its content. He says:

There’s a abstracted class wherever a website, oregon a publisher, oregon a quality enactment had explicitly said ‘do not scrape oregon crawl maine for immoderate different crushed than indexing maine truthful that different radical tin find this content.’ That’s a grey area, and I deliberation it’s going to enactment its mode done the courts.

But robots.txt is not a ineligible document. It, not just use, is the societal declaration that’s been with america since the ‘90s — and yet immoderate AI companies appear to beryllium ignoring it, too. Microsoft spouse OpenAI is reportedly among those ignoring it.

Read Entire Article