Twitch is using streamers' content to train Amazon's AI models unless users deliberately opt out. The setting is enabled by default and located in security settings. Streamers must manually disable it to prevent their VODs, clips and chat from being used to train generative AI. https://www.404media.co/twitch-training-amazon-ai-models-how-to-opt-out-setting/ #Media #SocialMedia #AI

YouTube is expanding access to its in-app AI chatbot, offering contextual responses across desktop, mobile and TV devices. The AI-powered function aims to enhance viewer engagement. https://www.socialmediatoday.com/news/youtube-expands-access-to-its-in-app-ai-chatbot/827757/ #Media #SocialMedia #AI

X is open sourcing its ranking algorithm and letting users check if they have been shadowbanned. The transparency tool will show if labels have been applied to accounts or posts. https://techcrunch.com/2026/08/13/x-open-sources-its-ranking-algorithm-letting-users-see-if-theyve-been-shadowbanned/ #Media #SocialMedia #DataPrivacy

A person has hidden prompt injection instructions in a US court filing, attempting to manipulate AI systems into ruling in their favour. The case highlights growing concerns about AI content moderation in legal settings. https://www.404media.co/person-hides-prompt-injection-in-legal-filing-telling-ai-to-side-with-them/ #Media #SocialMedia #AI

France's top court has blocked the government's plan to ban social media for under-15s, ruling it infringes on freedom of expression. President Macron says a new draft law will be prepared. https://bbc.co.uk/news/articles/cx2vj433xqlo #Media #SocialMedia #Policy

A new #study reveals that #onlinenegativity is not solely driven by attention-seeking algorithms. The study analysed over 2B #Reddit comments and found that the innate drive to differentiate oneself, or “semantic differentiation,” contributes to #negativity. As conversations progress, individuals seek to stand out by making unique comments, often leading to more negative discourse. https://news.harvard.edu/gazette/story/2026/08/does-vitriol-make-you-feel-special/?eicker.news #tech #news #ainews

US states are suing Meta in a landmark child privacy trial that could force Instagram and Facebook to overhaul their platforms for young users. The case may reshape how social media companies handle under-18 users. https://www.bbc.co.uk/news/articles/clyqpx6xk69o #Media #SocialMedia #Privacy

An expert witness hired by 3M used ChatGPT to write significant portions of his report in a 61 million USD lawsuit over an explosion that killed three people. Prompts asked the AI to show how 3M is 0% at fault. The case highlights how AI-generated content is entering courtrooms. https://www.404media.co/show-how-3m-is-0-at-fault-expert-witness-used-chatgpt-to-write-report-defending-company-in-deadly-explosion-lawsuit/ #Media #SocialMedia #AI

Meta's true AI costs exceed previous accounts with future obligations reaching nearly 700B USD in unreported expenditures, WSJ report reveals. https://www.socialmediatoday.com/news/metas-true-ai-costs-exceed-previous-accounts/828097/ #Media #SocialMedia #AI

Anthropic is changing how its AI writes text to embed invisible watermarks that can detect AI-generated content. The company will subtly alter word choices in a pattern only known to Anthropic, without adding hidden characters or metadata. The approach has sparked criticism that it prioritises detection over writing quality. https://www.404media.co/anthropics-text-watermarking-proves-ai-companies-do-not-care-at-all-about-writing/ #Media #SocialMedia #AI

X's algorithm prioritises engagement above all else and serves more ragebait to Democrat users, a new study has found. The research reveals a feedback loop where replying to inflammatory content leads to seeing more of it. https://www.404media.co/xs-algorithm-feeds-off-ragebait-and-impacts-democrats-more-study-finds/ #Media #SocialMedia #SocialNews

TikTok is adding AI-generated intros for live streams. The app will offer creators AI-powered summaries of themselves. This follows similar AI tools from competitors as platforms race to add generative features. https://www.socialmediatoday.com/news/tiktok-live-adds-ai-generated-intros/828220/ #Media #SocialMedia #AI

Amazon is buying and destroying rare books to train its AI systems, according to a new investigation. The report raises fresh questions about the ethics of AI training data and the preservation of cultural heritage. https://www.404media.co/podcast-amazon-is-destroying-rare-books-to-train-ai/ #Media #SocialMedia #AI

US senators are demanding TikTok explain why an internal experiment disabled safety features designed to prevent users being overwhelmed by harmful content. The platform wanted to test whether removing the safeguards made the app more engaging. The hearing comes amid ongoing scrutiny of the app impact on young users. https://techcrunch.com/2026/08/20/senators-demand-answers-from-tiktok-over-experiment-that-disabled-safeguards/ #Media #SocialMedia #ShortFormVideo

TikTok and ByteDance agree to pay 400M USD to settle US lawsuit alleging they collected data on millions of children under 13. The settlement is one of the largest child privacy cases involving a social media platform. https://www.bbc.co.uk/news/articles/cwyr0l45xjro #Media #SocialMedia #DataPrivacy

Fake AI-generated images of children in Cat in the Hat costumes have gone viral on TikTok, prompting police in Ireland and the UK to issue statements confirming the images are fabricated. The trend shows how quickly AI-generated content can spread and cause real concern. https://www.bbc.co.uk/news/articles/cwyqy29jn74o #Media #SocialMedia #TikTok

Meta faces a landmark trial where 29 US state attorneys general allege the company deliberately designed addictive social media systems. If found liable, Meta could face over 1 trillion USD in damages. https://www.socialmediatoday.com/news/will-meta-be-found-guilty-of-driving-social-media-addiction/828553/ #Media #SocialMedia #DataPrivacy

New Zealand is to introduce a bill banning children under 16 from social media. Platforms including TikTok, Instagram, Snapchat and Facebook could face fines of up to 10% of global revenue if they fail to verify users' age. The government says social media exposes children to harmful content and addictive technology. https://www.bbc.co.uk/news/articles/cj068q7qj19o #Media #SocialMedia #Privacy

Meta is developing an AI glasses kill switch to address privacy concerns. The patent filed in February describes a hardware mute switch that would disconnect all recording and AI functions. This comes as venues including UK restaurants, pubs and theatres are banning the glasses over privacy fears. https://www.socialmediatoday.com/news/meta-is-developing-an-ai-glasses-kill-switch/828667/ #Media #SocialMedia #AI

Israel is funding a synthetic think tank that has published over 100 articles in under a month, designed to influence AI chatbot responses. The Hanover Institute, run by US firm Piro Inc, publishes articles with question headlines about Israel and Palestine that appear engineered to tweak AI-generated answers. https://www.404media.co/israel-is-running-a-synthetic-think-tank-to-influence-ai-search-results/ #Media #SocialMedia #AI