med-mastodon.com is one of the many independent Mastodon servers you can use to participate in the fediverse.
Medical community on Mastodon

Administered by:

Server stats:

362
active users

#aitraining

0 posts0 participants0 posts today
Miguel Afonso Caetano<p>"For context, last week Facebook began showing users a prompt asking them to opt into "cloud processing," TechCrunch reported. Should you consent, this allows Facebook to grab stuff from your camera roll and upload it to Facebook's servers "on a regular basis" so it can generate recaps and "AI restylings" of your photos.</p><p>The important detail is that by opting in, Meta is asking you to agree to its AI terms, which state that, "once shared, you agree that Meta will analyze those images, including facial features, using AI." Meta would also gain the right to "retain and use" the information shared with its AI systems.</p><p>Your alarm bells should already be ringing. Any data that gets fed into an AI system runs the risk of being coughed up or reproduced in some shape or form. And asking for access to your entire camera roll so Meta's tech can "analyze" your photos is a huge and invasive escalation — it's shameless that Meta's even asking. Apparently, already using everyone's billions of Facebook and Instagram posts made since 2007 wasn't enough for Zuckerberg's tech juggernaut.</p><p>Moreover, Meta's AI terms don't make it clear if your unpublished camera roll photos it uses for "cloud processing" are safe from AI training. That's in stark contrast with the terms outlined for apps like Google Photos, the Verge noted, which explicitly state that your personal info won't be used as training data."</p><p><a href="https://futurism.com/meta-sketchy-training-ai-private-photos" rel="nofollow noopener" translate="no" target="_blank"><span class="invisible">https://</span><span class="ellipsis">futurism.com/meta-sketchy-trai</span><span class="invisible">ning-ai-private-photos</span></a></p><p><a href="https://tldr.nettime.org/tags/AI" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>AI</span></a> <a href="https://tldr.nettime.org/tags/GenerativeAI" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>GenerativeAI</span></a> <a href="https://tldr.nettime.org/tags/AITraining" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>AITraining</span></a> <a href="https://tldr.nettime.org/tags/Meta" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>Meta</span></a> <a href="https://tldr.nettime.org/tags/Privacy" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>Privacy</span></a> <a href="https://tldr.nettime.org/tags/DataProtection" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>DataProtection</span></a></p>
Ars Technica News<p>Pay up or stop scraping: Cloudflare program charges bots for each crawl <a href="https://arstechni.ca/MJGn" rel="nofollow noopener" translate="no" target="_blank"><span class="invisible">https://</span><span class="">arstechni.ca/MJGn</span><span class="invisible"></span></a> <a href="https://c.im/tags/ArtificialIntelligence" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>ArtificialIntelligence</span></a> <a href="https://c.im/tags/aiscraping" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>aiscraping</span></a> <a href="https://c.im/tags/AItraining" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>AItraining</span></a> <a href="https://c.im/tags/cloudflare" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>cloudflare</span></a> <a href="https://c.im/tags/robots" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>robots</span></a>.txt <a href="https://c.im/tags/aicrawler" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>aicrawler</span></a> <a href="https://c.im/tags/Policy" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>Policy</span></a> <a href="https://c.im/tags/aibots" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>aibots</span></a> <a href="https://c.im/tags/AI" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>AI</span></a></p>
Inautilo<p><a href="https://mastodon.social/tags/Design" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>Design</span></a> <a href="https://mastodon.social/tags/Proposals" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>Proposals</span></a><br>CC Signals · A new way to let creators set fair AI use terms <a href="https://ilo.im/164yrv" rel="nofollow noopener" translate="no" target="_blank"><span class="invisible">https://</span><span class="">ilo.im/164yrv</span><span class="invisible"></span></a></p><p>_____<br><a href="https://mastodon.social/tags/Business" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>Business</span></a> <a href="https://mastodon.social/tags/AI" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>AI</span></a> <a href="https://mastodon.social/tags/AiTraining" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>AiTraining</span></a> <a href="https://mastodon.social/tags/Responsibility" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>Responsibility</span></a> <a href="https://mastodon.social/tags/CreativeCommons" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>CreativeCommons</span></a> <a href="https://mastodon.social/tags/Creativity" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>Creativity</span></a> <a href="https://mastodon.social/tags/Copyright" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>Copyright</span></a> <a href="https://mastodon.social/tags/Content" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>Content</span></a> <a href="https://mastodon.social/tags/Website" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>Website</span></a> <a href="https://mastodon.social/tags/Blog" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>Blog</span></a></p>
Miguel Afonso Caetano<p>"Much of the Kadrey ruling by Judge Vince Chhabria is dicta—meaning, the opinion spends many paragraphs on what it thinks could justify ruling in favor of the author plaintiffs, if only they had managed to present different facts (rather than pure speculation). The court then rules in Meta’s favor because the plaintiffs only offered speculation. </p><p>But it makes a number of errors along the way to the right outcome. At the top, the ruling broadly proclaims that training AI without buying a license to use each and every piece of copyrighted training material will be “illegal” in “most cases.” The court asserted that fair use usually won’t apply to AI training uses even though training is a “highly transformative” process, because of hypothetical “market dilution” scenarios where competition from AI-generated works could reduce the value of the books used to train the AI model..</p><p>That theory, in turn, depends on three mistaken premises. First, that the most important factor for determining fair use is whether the use might cause market harm. That’s not correct. Since its seminal 1994 opinion in Cambell v Acuff-Rose, the Supreme Court has been very clear that no single factor controls the fair use analysis.</p><p>Second, that an AI developer would typically seek to train a model entirely on a certain type of work, and then use that model to generate new works in the exact same genre, which would then compete with the works on which it was trained, such that the market for the original works is harmed. As the Kadrey ruling notes, there was no evidence that Llama was intended to to, or does, anything like that, nor will most LLMs for the exact reasons discussed in Bartz.</p><p>Third, as a matter of law, copyright doesn't prevent “market dilution” unless the new works are otherwise infringing. In fact, the whole purpose of copyright is to be an engine for new expression."</p><p><a href="https://www.eff.org/deeplinks/2025/06/two-courts-rule-generative-ai-and-fair-use-one-gets-it-right" rel="nofollow noopener" translate="no" target="_blank"><span class="invisible">https://www.</span><span class="ellipsis">eff.org/deeplinks/2025/06/two-</span><span class="invisible">courts-rule-generative-ai-and-fair-use-one-gets-it-right</span></a></p><p><a href="https://tldr.nettime.org/tags/AI" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>AI</span></a> <a href="https://tldr.nettime.org/tags/GenerativeAI" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>GenerativeAI</span></a> <a href="https://tldr.nettime.org/tags/AITraining" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>AITraining</span></a> <a href="https://tldr.nettime.org/tags/FairUse" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>FairUse</span></a> <a href="https://tldr.nettime.org/tags/Copyright" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>Copyright</span></a> <a href="https://tldr.nettime.org/tags/IP" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>IP</span></a> <a href="https://tldr.nettime.org/tags/Meta" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>Meta</span></a> <a href="https://tldr.nettime.org/tags/Chatbots" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>Chatbots</span></a> <a href="https://tldr.nettime.org/tags/LLMs" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>LLMs</span></a></p>
Ars Technica News<p>Book authors made the wrong arguments in Meta AI training case, judge says <a href="https://arstechni.ca/8nfR" rel="nofollow noopener" translate="no" target="_blank"><span class="invisible">https://</span><span class="">arstechni.ca/8nfR</span><span class="invisible"></span></a> <a href="https://c.im/tags/copyrightinfringement" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>copyrightinfringement</span></a> <a href="https://c.im/tags/AItraining" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>AItraining</span></a> <a href="https://c.im/tags/torrenting" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>torrenting</span></a> <a href="https://c.im/tags/copyright" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>copyright</span></a> <a href="https://c.im/tags/leeching" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>leeching</span></a> <a href="https://c.im/tags/Policy" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>Policy</span></a> <a href="https://c.im/tags/LLaMA" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>LLaMA</span></a> <a href="https://c.im/tags/meta" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>meta</span></a> <a href="https://c.im/tags/AI" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>AI</span></a></p>
Yonhap Infomax News<p>Meta Platforms Inc. won a key U.S. court case allowing the use of books for AI model training under fair use, a decision with limited scope that intensifies the debate over copyright in artificial intelligence development.<br><a href="https://mastodon.social/tags/YonhapInfomax" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>YonhapInfomax</span></a> <a href="https://mastodon.social/tags/MetaPlatforms" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>MetaPlatforms</span></a> <a href="https://mastodon.social/tags/CopyrightLawsuit" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>CopyrightLawsuit</span></a> <a href="https://mastodon.social/tags/AITraining" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>AITraining</span></a> <a href="https://mastodon.social/tags/Llama" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>Llama</span></a> <a href="https://mastodon.social/tags/FairUse" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>FairUse</span></a> <a href="https://mastodon.social/tags/Economics" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>Economics</span></a> <a href="https://mastodon.social/tags/FinancialMarkets" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>FinancialMarkets</span></a> <a href="https://mastodon.social/tags/Banking" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>Banking</span></a> <a href="https://mastodon.social/tags/Securities" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>Securities</span></a> <a href="https://mastodon.social/tags/Bonds" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>Bonds</span></a> <a href="https://mastodon.social/tags/StockMarket" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>StockMarket</span></a> <br><a href="https://en.infomaxai.com/news/articleView.html?idxno=69484" rel="nofollow noopener" translate="no" target="_blank"><span class="invisible">https://</span><span class="ellipsis">en.infomaxai.com/news/articleV</span><span class="invisible">iew.html?idxno=69484</span></a></p>
Ars Technica News<p>Anthropic destroyed millions of print books to build its AI models <a href="https://arstechni.ca/YLWE" rel="nofollow noopener" translate="no" target="_blank"><span class="invisible">https://</span><span class="">arstechni.ca/YLWE</span><span class="invisible"></span></a> <a href="https://c.im/tags/InternetArchive" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>InternetArchive</span></a> <a href="https://c.im/tags/machinelearning" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>machinelearning</span></a> <a href="https://c.im/tags/AIdevelopment" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>AIdevelopment</span></a> <a href="https://c.im/tags/bookscanning" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>bookscanning</span></a> <a href="https://c.im/tags/legalrulings" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>legalrulings</span></a> <a href="https://c.im/tags/trainingdata" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>trainingdata</span></a> <a href="https://c.im/tags/AIcompanies" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>AIcompanies</span></a> <a href="https://c.im/tags/googlebooks" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>googlebooks</span></a> <a href="https://c.im/tags/AIresearch" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>AIresearch</span></a> <a href="https://c.im/tags/AItraining" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>AItraining</span></a> <a href="https://c.im/tags/Anthropic" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>Anthropic</span></a> <a href="https://c.im/tags/copyright" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>copyright</span></a> <a href="https://c.im/tags/AIethics" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>AIethics</span></a> <a href="https://c.im/tags/scanning" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>scanning</span></a> <a href="https://c.im/tags/fairuse" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>fairuse</span></a> <a href="https://c.im/tags/Biz" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>Biz</span></a>&amp;IT <a href="https://c.im/tags/Policy" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>Policy</span></a> <a href="https://c.im/tags/Claude" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>Claude</span></a> <a href="https://c.im/tags/AIlaw" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>AIlaw</span></a> <a href="https://c.im/tags/AI" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>AI</span></a></p>
Mastokarl 🇺🇦<p>Told you so 😜</p><p><a href="https://www.theverge.com/news/692015/anthropic-wins-a-major-fair-use-victory-for-ai-but-its-still-in-trouble-for-stealing-books" rel="nofollow noopener" translate="no" target="_blank"><span class="invisible">https://www.</span><span class="ellipsis">theverge.com/news/692015/anthr</span><span class="invisible">opic-wins-a-major-fair-use-victory-for-ai-but-its-still-in-trouble-for-stealing-books</span></a></p><p><a href="https://mastodon.social/tags/aiTraining" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>aiTraining</span></a> <a href="https://mastodon.social/tags/fairUse" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>fairUse</span></a> <a href="https://mastodon.social/tags/learningIsNotStealing" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>learningIsNotStealing</span></a></p>
Miguel Afonso Caetano<p>"A new paper from researchers at Stanford, Cornell, and West Virginia University seems to show that one version of Meta’s flagship AI model, Llama 3.1, has memorized almost the whole of the first Harry Potter book. This finding could have far-reaching copyright implications for the AI industry and impact authors and creatives who are already part of class-action lawsuits against Meta. </p><p>Researchers tested a bunch of different widely-available free large language models to see what percentage of 56 different books they could reproduce. The researchers fed the models hundreds of short text snippets from those books and measured how well it could recite the next lines. The titles were a random sampling of popular, lesser-known, and public domain works drawn from the now-defunct and controversial Books3 dataset that Meta used to train its models, as well as books by plaintiffs in the recent, and ongoing, Kadrey vs Meta class-action lawsuit. </p><p>According to Mark A. Lemley, one of the study authors, this finding might have some interesting implications. AI companies argue that their models are generative—as in, they make new stuff, rather than just being fancy search engines. On the other hand, authors and news outlets are suing on the basis that AI is just remixing existing material, including copyrighted content. “I think what we show in the paper is that neither of those characterizations is accurate,” says Lemley."</p><p><a href="https://www.404media.co/meta-ai-model-memorized-harry-potter-books/" rel="nofollow noopener" translate="no" target="_blank"><span class="invisible">https://www.</span><span class="ellipsis">404media.co/meta-ai-model-memo</span><span class="invisible">rized-harry-potter-books/</span></a></p><p><a href="https://tldr.nettime.org/tags/AI" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>AI</span></a> <a href="https://tldr.nettime.org/tags/GenerativeAI" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>GenerativeAI</span></a> <a href="https://tldr.nettime.org/tags/Meta" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>Meta</span></a> <a href="https://tldr.nettime.org/tags/Copyright" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>Copyright</span></a> <a href="https://tldr.nettime.org/tags/AITraining" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>AITraining</span></a> <a href="https://tldr.nettime.org/tags/FairUse" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>FairUse</span></a></p>
Miguel Afonso Caetano<p>"A federal judge in San Francisco ruled late on Monday that Anthropic's use of books without permission to train its artificial intelligence system was legal under U.S. copyright law.</p><p>Siding with tech companies on a pivotal question for the AI industry, U.S. District Judge William Alsup said Anthropic made "fair use" of books by writers Andrea Bartz, Charles Graeber and Kirk Wallace Johnson to train its Claude large language model.</p><p>Alsup also said, however, that Anthropic's copying and storage of more than 7 million pirated books in a "central library" infringed the authors' copyrights and was not fair use. The judge has ordered a trial in December to determine how much Anthropic owes for the infringement.</p><p>U.S. copyright law says that willful copyright infringement can justify statutory damages of up to $150,000 per work."</p><p><a href="https://www.reuters.com/legal/litigation/anthropic-wins-key-ruling-ai-authors-copyright-lawsuit-2025-06-24/" rel="nofollow noopener" translate="no" target="_blank"><span class="invisible">https://www.</span><span class="ellipsis">reuters.com/legal/litigation/a</span><span class="invisible">nthropic-wins-key-ruling-ai-authors-copyright-lawsuit-2025-06-24/</span></a></p><p><a href="https://tldr.nettime.org/tags/AI" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>AI</span></a> <a href="https://tldr.nettime.org/tags/GenerativeAI" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>GenerativeAI</span></a> <a href="https://tldr.nettime.org/tags/USA" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>USA</span></a> <a href="https://tldr.nettime.org/tags/Copyright" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>Copyright</span></a> <a href="https://tldr.nettime.org/tags/Anthropic" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>Anthropic</span></a> <a href="https://tldr.nettime.org/tags/FairUse" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>FairUse</span></a> <a href="https://tldr.nettime.org/tags/Claude" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>Claude</span></a> <a href="https://tldr.nettime.org/tags/AITraining" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>AITraining</span></a></p>
BGDon<p>Chock up a win for Anthropic. </p><p>Judge rules Anthropic's use of books without permission to train its artificial intelligence system was legal under U.S. copyright law, accepting the position put forward by Anthropic that it made fair use of the books and that U.S. copyright law "not only allows, but encourages" its AI training because it promotes human creativity. </p><p><a href="https://tech.yahoo.com/ai/articles/anthropic-wins-key-ruling-ai-133440779.html" rel="nofollow noopener" translate="no" target="_blank"><span class="invisible">https://</span><span class="ellipsis">tech.yahoo.com/ai/articles/ant</span><span class="invisible">hropic-wins-key-ruling-ai-133440779.html</span></a> <a href="https://techhub.social/tags/AI" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>AI</span></a> <a href="https://techhub.social/tags/AITraining" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>AITraining</span></a> <a href="https://techhub.social/tags/Anthropic" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>Anthropic</span></a> <a href="https://techhub.social/tags/LLMs" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>LLMs</span></a> <a href="https://techhub.social/tags/Copyright" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>Copyright</span></a> <a href="https://techhub.social/tags/Lawsuit" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>Lawsuit</span></a> <a href="https://techhub.social/tags/Claude" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>Claude</span></a></p>
Miguel Afonso Caetano<p>"If you stumbled across Terence Broad’s AI-generated artwork (un)stable equilibrium on YouTube, you might assume he’d trained a model on the works of the painter Mark Rothko — the earlier, lighter pieces, before his vision became darker and suffused with doom. Like early-period Rothko, Broad’s AI-generated images consist of simple fields of pure color, but they’re morphing, continuously changing form and hue.</p><p>But Broad didn’t train his AI on Rothko; he didn’t train it on any data at all. By hacking a neural network, and locking elements of it into a recursive loop, he was able to induce this AI into producing images without any training data at all — no inputs, no influences. Depending on your perspective, Broad’s art is either a pioneering display of pure artificial creativity, a look into the very soul of AI, or a clever but meaningless electronic by-product, closer to guitar feedback than music. In any case, his work points the way toward a more creative and ethical use of generative AI beyond the large-scale manufacture of derivative slop now oozing through our visual culture.</p><p>Broad has deep reservations about the ethics of training generative AI on other people’s work, but his main inspiration for (un)stable equilibrium wasn’t philosophical; it was a crappy job."</p><p><a href="https://www.theverge.com/ai-artificial-intelligence/688576/feed-ai-nothing" rel="nofollow noopener" translate="no" target="_blank"><span class="invisible">https://www.</span><span class="ellipsis">theverge.com/ai-artificial-int</span><span class="invisible">elligence/688576/feed-ai-nothing</span></a></p><p><a href="https://tldr.nettime.org/tags/AI" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>AI</span></a> <a href="https://tldr.nettime.org/tags/GenerativeAI" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>GenerativeAI</span></a> <a href="https://tldr.nettime.org/tags/GeneratedImages" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>GeneratedImages</span></a> <a href="https://tldr.nettime.org/tags/AITraining" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>AITraining</span></a> <a href="https://tldr.nettime.org/tags/Copyright" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>Copyright</span></a> <a href="https://tldr.nettime.org/tags/IP" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>IP</span></a></p>
The Tired Horizon<p><a href="https://mstdn.social/tags/PSA" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>PSA</span></a> please retoot for a wider awareness</p><p>DO NOT UPLOAD MUSIC TO SOUNDCLOUD!!</p><p>Soundcloud now freezes accounts so they cannot be deleted. This appears to have been made policy after the backlash of their AI policy.</p><p>Any attempt to get this rectified sends you around the bend in their AI chatbot that just spits back garbage. (Which is what I am going through now)<br><a href="https://mstdn.social/tags/soundcloud" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>soundcloud</span></a> <a href="https://mstdn.social/tags/music" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>music</span></a> <a href="https://mstdn.social/tags/streaming" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>streaming</span></a> <a href="https://mstdn.social/tags/AI" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>AI</span></a> <a href="https://mstdn.social/tags/AItraining" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>AItraining</span></a> <a href="https://mstdn.social/tags/aimusic" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>aimusic</span></a> <a href="https://mstdn.social/tags/tech" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>tech</span></a></p>
Miguel Afonso Caetano<p>A for-profit corporation that makes money out of other users' content is accusing another company of trying to do exactly the same. How amusing can that be? :-D </p><p>"Reddit said the AI company unlawfully used Reddit’s data for commercial purposes without paying for it and without abiding by the company’s user data policy, according to the complaint, which was filed Wednesday in California.</p><p>“Anthropic is in fact intentionally trained on the personal data of Reddit users without ever requesting their consent,” the complaint says, alleging that Anthropic’s conduct runs counter to how it “bills itself as the white knight of the AI industry.”</p><p>Reddit, the online discussion forum where users can post anonymously and ask each other questions, has reached formal agreements with both OpenAI and Google to license Reddit’s valuable human user data. </p><p>Anthropic didn’t immediately comment."</p><p><a href="https://www.wsj.com/tech/ai/reddit-lawsuit-anthropic-ai-3b9624dd" rel="nofollow noopener" translate="no" target="_blank"><span class="invisible">https://www.</span><span class="ellipsis">wsj.com/tech/ai/reddit-lawsuit</span><span class="invisible">-anthropic-ai-3b9624dd</span></a></p><p><a href="https://tldr.nettime.org/tags/AI" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>AI</span></a> <a href="https://tldr.nettime.org/tags/GenerativeAI" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>GenerativeAI</span></a> <a href="https://tldr.nettime.org/tags/Reddit" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>Reddit</span></a> <a href="https://tldr.nettime.org/tags/SocialMedia" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>SocialMedia</span></a> <a href="https://tldr.nettime.org/tags/AITraining" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>AITraining</span></a> <a href="https://tldr.nettime.org/tags/Anthropic" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>Anthropic</span></a> <a href="https://tldr.nettime.org/tags/Claude" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>Claude</span></a></p>
Miguel Afonso Caetano<p>"Ai2 tested DataDecide across a wide range of datasets and model sizes, using 10 benchmarks to evaluate how well small models predict large-scale performance. The findings aren’t earth-shattering, but they present useful takeaways for AI developers and researchers.</p><p>For one, Ai2 found that small models (around 150 million parameters) can predict large-scale outcomes with surprising accuracy. Some benchmarks reached over 80% decision accuracy using just 0.01% of the compute compared to billion-parameter models.</p><p>Since small-model experiments use less compute than other methods, developers don’t need to run full-scale tests just to predict outcomes. “The promise of this work is lower compute costs during training,” said Pijanowski.</p><p>Ai2 found that scaling laws didn’t outperform the simpler method of ranking datasets by small-model results. Scaling laws, a more sophisticated and more costly testing method, aim to predict how accuracy improves with model size. For now, “just stick with ablating things at one scale,” advised Magnusson.</p><p>The findings should give LLM devs pause for thought, Hunt said: “There are scaling laws that have been derived from empirical studies between data volume, compute resources and performance. Ai2’s research points out that we may want to revisit some of those assumptions.”"</p><p><a href="https://thenewstack.io/new-tools-help-llm-developers-choose-better-pre-training-data/" rel="nofollow noopener" translate="no" target="_blank"><span class="invisible">https://</span><span class="ellipsis">thenewstack.io/new-tools-help-</span><span class="invisible">llm-developers-choose-better-pre-training-data/</span></a></p><p><a href="https://tldr.nettime.org/tags/AI" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>AI</span></a> <a href="https://tldr.nettime.org/tags/GenerativeAI" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>GenerativeAI</span></a> <a href="https://tldr.nettime.org/tags/LLMs" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>LLMs</span></a> <a href="https://tldr.nettime.org/tags/AITraining" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>AITraining</span></a> <a href="https://tldr.nettime.org/tags/SLMs" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>SLMs</span></a></p>
Tuckers Nuts Resist! 🇺🇦 <p><span class="h-card" translate="no"><a href="https://beige.party/@maxleibman" class="u-url mention" rel="nofollow noopener" target="_blank">@<span>maxleibman</span></a></span> <br>Samesies, Max!<br>Like Nick Clegg, Eye believe that being required to obtain consent from rights holders before watching copyrighted streaming movies and shows would “kill” my enjoyment of the streaming content industry.<br><a href="https://mstdn.social/tags/NickClegg" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>NickClegg</span></a> <a href="https://mstdn.social/tags/AiTraining" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>AiTraining</span></a> <a href="https://mstdn.social/tags/TuckersBalls" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>TuckersBalls</span></a></p>
Miguel Afonso Caetano<p>"The number of Google contractors working on various projects isn’t public knowledge, but the company may rely on as many as 12,000 AI workers across seven to 10 different contractors, Wait, who started working with GlobalLogic raters after they reached out to her last February, estimates. Other artificial intelligence engines depend on thousands more. At GlobalLogic, which declined to comment for this story, workers are assigned to support Google engineers working on projects that include Gemini, Google’s AI “assistant.”</p><p>In some ways, this type of work is not new—engines like Google have long relied on underpaid raters to train their search algorithms—but the AI boom has led to exponential unregulated growth in this workforce, as tech juggernauts pour billions of dollars into the race to capture market share. In 2023, a Time investigation found that OpenAI had paid Kenyan contractors less than $2 an hour to watch and identify violent and abusive content. The workers, who also rated content for Meta, lost their jobs when they tried to organize. A lawsuit against Meta is ongoing. On April 30, moderators launched a global trade union alliance across nine countries; so far, the United States is not one of them.</p><p>GlobalLogic workers, who are based in the United States, do make more money than their Kenyan peers, with wages for generalist raters starting at $16 an hour. So-called “super raters” are paid $20 or more an hour to do the same kind of work, because they usually have master’s or doctoral degrees—although that’s well under the average for American workers with comparable qualifications. In other ways, though, the work can be much the same across national borders. Rachael Sawyer works on a project training Google’s AI to filter out hateful and violent content, including child sexual abuse material."</p><p><a href="https://www.thenation.com/article/society/google-ai-workers-union/" rel="nofollow noopener" translate="no" target="_blank"><span class="invisible">https://www.</span><span class="ellipsis">thenation.com/article/society/</span><span class="invisible">google-ai-workers-union/</span></a></p><p><a href="https://tldr.nettime.org/tags/AI" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>AI</span></a> <a href="https://tldr.nettime.org/tags/GenerativeAI" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>GenerativeAI</span></a> <a href="https://tldr.nettime.org/tags/Google" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>Google</span></a> <a href="https://tldr.nettime.org/tags/AITraining" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>AITraining</span></a> <a href="https://tldr.nettime.org/tags/Gemini" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>Gemini</span></a> <a href="https://tldr.nettime.org/tags/GlobalLogic" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>GlobalLogic</span></a> <a href="https://tldr.nettime.org/tags/Hitachi" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>Hitachi</span></a></p>
ResearchBuzz: Firehose<p>AFP: German court says Meta can use user data to train AI. “A German court on Friday dismissed an injunction request brought by consumer protection groups to prevent US tech giant Meta from using user data from Facebook and Instagram to train artificial intelligence systems.”</p><p><a href="https://rbfirehose.com/2025/05/28/afp-german-court-says-meta-can-use-user-data-to-train-ai/" class="" rel="nofollow noopener" target="_blank">https://rbfirehose.com/2025/05/28/afp-german-court-says-meta-can-use-user-data-to-train-ai/</a></p>
Miguel Afonso Caetano<p>I find it astonishing how authors are so afraid that AI is going to steal their readers. If AI only generates derivative works, why are you, as a creator and the only entity capable of producing truly creative works so afraid? </p><p>Unless you believe in the fiction that you can own artificial property. In reality, once you let a work out in the world, you can't possibly expect to "own" it and control its distribution - unless you want to enforce a totalitarian dictatorship. So, I repeat, what are you so afraid of? </p><p>Unless authors start to understand that creative works can't be really protected against unauthorized copying and distribution and that copyright is a monopoly granted by States, they will continue to repeat the same mistakes, will depriving the public of access to knowledge and culture. </p><p>"In late 2024, we surveyed over 400 members of the Australian Society of Authors, the national peak body for writers and illustrators. We asked about their use of AI, their understanding of how generative models are trained, and whether they would agree to their work being used for training – with or without compensation.</p><p>79% said they would not allow their existing work to be used to train AI models, even if they were paid. Almost as many – 77% – said the same about future work.</p><p>Among those open to payment, half expected at least $A1,000 per work. A small number nominated figures in the tens or hundreds of thousands.</p><p>But the dominant response, from both established and emerging authors, was a firm “no”.</p><p>This presents a serious roadblock for those hoping publishers might broker blanket licensing agreements with AI firms. If most authors are unwilling to grant permission under any terms, then standard contract clauses or opt-in models are unlikely to deliver a practical or ethical solution."</p><p><a href="https://theconversation.com/new-research-reveals-australian-authors-say-no-to-ai-using-their-work-even-if-money-is-on-the-table-257243" rel="nofollow noopener" translate="no" target="_blank"><span class="invisible">https://</span><span class="ellipsis">theconversation.com/new-resear</span><span class="invisible">ch-reveals-australian-authors-say-no-to-ai-using-their-work-even-if-money-is-on-the-table-257243</span></a></p><p><a href="https://tldr.nettime.org/tags/Australia" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>Australia</span></a> <a href="https://tldr.nettime.org/tags/AI" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>AI</span></a> <a href="https://tldr.nettime.org/tags/GenerativeAI" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>GenerativeAI</span></a> <a href="https://tldr.nettime.org/tags/AITraining" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>AITraining</span></a> <a href="https://tldr.nettime.org/tags/Copyright" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>Copyright</span></a> <a href="https://tldr.nettime.org/tags/RentSeeking" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>RentSeeking</span></a> <a href="https://tldr.nettime.org/tags/Monopolies" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>Monopolies</span></a></p>
Miguel Afonso Caetano<p>COMPLETE TRAVESTY: Creativity is not an industry. Anyone who dares to say that knows nothing about art, culture, creativity, and manufacturing. You cannot manufacture creativity. Either a work is considered creative by the public and the critics or not.</p><p>Am I stealing your words in favor of a rent-seeking, money-grabbing little scheme by copying and pasting it here? Do I have to pay you a license for your fake, artificial property? And you consider yourself a representative of artists and authors? </p><p>"My colleagues and I from all sides in the House of Lords have acted where the government has refused, adding emergency transparency measures to the legislation – the data (use and access) bill – that is passing through parliament. Our amendment would allow existing copyright law to be enforced: copyright owners would understand when, where and by whom their work was being stolen to train AI. The logic being that if an AI firm has to disclose evidence of theft, it will not steal in the first place. These measures, voted for in ever-increasing numbers by lords from all parties – and notable grandees from the government’s own backbenches – were voted down by a government wielding its significant, if reluctant, majority."</p><p><a href="https://www.theguardian.com/commentisfree/2025/may/24/ai-britain-creative-industries-government-data-bill" rel="nofollow noopener" translate="no" target="_blank"><span class="invisible">https://www.</span><span class="ellipsis">theguardian.com/commentisfree/</span><span class="invisible">2025/may/24/ai-britain-creative-industries-government-data-bill</span></a></p><p><a href="https://tldr.nettime.org/tags/UK" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>UK</span></a> <a href="https://tldr.nettime.org/tags/RentSeeking" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>RentSeeking</span></a> <a href="https://tldr.nettime.org/tags/AI" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>AI</span></a> <a href="https://tldr.nettime.org/tags/GenerativeAI" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>GenerativeAI</span></a> <a href="https://tldr.nettime.org/tags/MoneyGrabbing" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>MoneyGrabbing</span></a> <a href="https://tldr.nettime.org/tags/Feudalism" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>Feudalism</span></a> <a href="https://tldr.nettime.org/tags/Monopolies" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>Monopolies</span></a> <a href="https://tldr.nettime.org/tags/IP" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>IP</span></a> <a href="https://tldr.nettime.org/tags/Copyright" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>Copyright</span></a> <a href="https://tldr.nettime.org/tags/AITraining" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>AITraining</span></a></p>