Training AI? Scan and destroy books. Everyone else is doing it.

Even big tech hates AI slop writing.Anthropic took this to heart, alongside other big tech firms, running to libraries to collect hundreds of thousands of pre-2022 (AKA, pre-ChatGPT AI slop text) books to scan to train their AI models. They’re using a method pioneered by Google back in the day…spoiler alert, the books don’t make it.Sound concerning?We thought so, but the court thinks so less: the recent ruling and settlement with Anthropic using scanned copies of books (and then destroying the originals) to train their AI models suggests this is a big win for big tech, and a big loss for human authors.Project Panama was “our effort to destructively scan all the books in the world.” Quote: Anthropic. Meanwhile, this method of using books to train AI models was not against the Copyright Act, according to a judge, who ruled this use of books as “quintessentially transformative.”This isn’t the first time big tech has scanned and shredded books, or used LibGen or other sites to scrape human-written texts.What happened, and why is this concerning? Listen to the full episode to find out.00:00 A win for big tech, and a loss for authors: we summarize Project Panama (Anthropic’s plan to “destructively scan all the books in the world”), the recent settlement / $1.5 billion payout to authors for LibGen vs. the Copyright Act07:42 Details on the recently-settled class action suit (Bartz vs. Anthropic): what happened, and why are the results so concerning?11:45 Meta was sued last year and won against authors for using pirated copies of their novels from LibGen to train their AI models - what gives?13:13 Internal turmoil: employees vs. executives on whether these book-scanning AI-training practices are okay14:11 Setting a precedent: What do Meta and Anthropic’s cases mean for IP, authors, novels, academic texts, and any related writing behind a paywall? Health GPT is discussed15:34 Ah, the precedent rears its head. Up next: big tech is rolling out contracts with media houses including news outlets, publishing firms, and individual authors to use their written content for training AI models, preemptively. Consent and compensation are good, right?17:56 Oh, the irony: the AI slop wheel will continue, despite big tech hoping to avoid it by buying all these pre-2022 (pre-ChatGPT) books written by real human authors to train their AI models. We also debate self-expression, moral outrage against AI slop writing, and big tech’s hope for writers to jump back on the hamster wheel.22:42 Where will we get good data, then, amongst the pile of AI slop? Not to fear, there’s another source of untainted human training data: Meta’s AI glasses! Real world data! Yay! Who cares about consent!31:17 Google next? Let’s bring Tom Turvey back in! A new lawsuit out against Google for using scanned book copies meant for Google Books to train their AI models. And…that’s a wrap.-This is Our Lives With Bots, the show where we ask important, timely questions about what it means to live with our bot counterparts. From time to time, we also dive deep into what an AI future might look like for us. Sometimes we agree, sometimes we spiral, but we always go deep.Rose and Angy are psychologists with degrees in psychology, artificial intelligence, and ethics. They have conducted research in human-AI interaction and created this podcast to make information about AI accessible to you. Rose holds a PhD in psychology and social policy and is a postdoctoral researcher in computer science with Arvind Narayanan. Angy holds a PhD in psychology and a masters in AI Ethics and is currently a CXO. You can learn more about us at ⁠ourliveswithbots.com⁠.-Links:Bartz vs. Anthropic lawsuit related to Project Panama: Euro News, NPRWhat judge said in a summary rulingProject Panama: Washington PostGoogle lawsuit for Gemini: The Guardian