TikTok will flag users who upload artificial intelligence-generated content (AIGC) to the video-sharing site from other platforms, the company says, becoming the first big video site to automatically label such content for users to see. Content created using TikTok's own AI tools is already automatically marked as such to viewers, and the company has required creators to manually add the same labels to their own content, but until now they have been able to evade the rules and pass off generated material as authentic by uploading it from other platforms. Now, the company will begin using digital watermarks created by the cross-industry group Coalition for Content Provenance and Authenticity (C2PA) to identify and label as much AIGC as it can. "AI enables incredible creative opportunities but can confuse or mislead viewers if they don't know content was AI-generated," said Adam Presser, the head of operations and trust and safety at TikTok. "Labelling helps make that context clear – which is why we label AIGC made with TikTok AI effects, and have required creators to label realistic AIGC for over a year."

After OpenAI released GPT-3 in July 2020, independent artificial intelligence researcher Shawn Presser and a few of his fellow machine-learning enthusiasts set a challenge for themselves: Could they recreate it? "We were like, OK, there's actually not that much standing in the way of us doing this ourselves," Presser says. So what if OpenAI had deep pockets and a head start? That summer, they pored over papers about GPT-3, strategizing in marathon Discord chats about how to best approximate its training data sets. Presser honed in on the books they needed.

One of the most troubling issues around generative AI is simple: It's being made in secret. To produce humanlike answers to questions, systems such as ChatGPT process huge quantities of written material. But few people outside of companies such as Meta and OpenAI know the full extent of the texts these programs have been trained on. Some training text comes from Wikipedia and other online writing, but high-quality generative AI requires higher-quality input than is usually found on the internet--that is, it requires the kind found in books. But neither the lawsuit itself nor the commentary surrounding it has offered a look under the hood: We have not previously known for certain whether LLaMA was trained on Silverman's, Kadrey's, or Golden's books, or any others, for that matter.

The world's most exciting technology show, CES, descends upon Las Vegas this week with more than 100,000 industry professionals to network with, hundreds of sessions to attend, and two million net square feet of expo space to explore. Walk through the home of the future. Have a robot mix you a cocktail. Party with Imagine Dragons, DJ Snoopadelic, Def Leppard, Tipper Gore and Paris Hilton. This guide will show you what's happening where.