The Slop Report - June 10, 2026
Your daily digest of AI-generated content news from around the web. All signal, no slop.
1. Bank of England warns of AI scams as deepfakes of Farage-Bailey fight spread
The Guardian Tech - · Jun 9
Bank of England Governor Andrew Bailey has issued a public warning after deepfake videos circulated on X showing him fighting with Reform UK leader Nigel Farage on BBC’s Question Time, some depicting Farage with a gun. Bailey used the incident to alert the public to the rising threat of AI-generated scams impersonating central banks and public figures, urging people to report fraudulent content and remain vigilant online. The warning highlights how increasingly sophisticated AI technology is being exploited for fraud, particularly targeting vulnerable people through fake advertisements and impersonations.
2. Chinese activist in UK told by X that abusive deepfakes do not breach rules
The Guardian Tech - · Jun 10
Summary Chinese activist Apple Peiqing Ni, who leads the UK-based China Dissent Network, was
targeted by deepfake posts on X portraying her as a drug addict and sexually promiscuous individual after posting about a Tiananmen Square commemoration. X’s automated systems initially rejected her complaints, ruling the abusive content did not violate platform rules on harassment, despite explicit prohibitions against malicious targeting—the account was only suspended after The Guardian contacted X’s press office. The case highlights failures in X’s content moderation systems and raises concerns about coordinated harassment campaigns targeting dissidents, particularly those criticizing China.
3. ChatGPT is recommending scam websites that will steal your credit card info
Digital Trends - · Jun 10
Summary ChatGPT is directing users to fraudulent retail websites designed to steal credit card
information, particularly sites impersonating defunct brands like Russell & Bromley (a British footwear retailer that closed in January 2026). Security firm Ask Silver discovered the scam, with researchers suggesting fraudsters may have “poisoned” ChatGPT’s training data to boost fake sites in search results. This vulnerability is concerning as AI becomes increasingly integrated into the shopping process, and experts warn the problem will likely grow without swift action.
4. Show HN: Eatmydata.ai – Local-First Question-to-SQL-to-Dashboard AI
Hacker News - · Jun 10
I don’t see an actual news story or link provided in your message—just the text “eatmydata,” which appears to be a Linux command or tool name rather than a news article. Could you please share the news story you’d like me to summarize? You can either paste the article text, provide a headline and excerpt, or share the URL.
5. Ambani-backed Addverb wants $100 million to build India’s answer to Unitree and Tesla Optimus
The Next Web - · Jun 10
Addverb Technologies, an Indian robotics startup backed by Mukesh Ambani’s Reliance Industries, is seeking over $100 million to develop humanoid and quadruped robots and AI training systems as it aims to crack the top 10 globally within five years. The company currently manufactures industrial robots for logistics and warehouse automation, projects $136 million in revenue this year, but hasn’t yet achieved net profitability and ranks outside the global top 30 in robotics market share. The funding matters because Addverb is betting that vertical integration and domestic technology development can help it compete against established Japanese and European manufacturers, well-funded Chinese competitors, and Tesla’s robotics programs.
6. Meta signs first AI data center deal in India with Reliance
TechCrunch AI - · Jun 10
Meta has partnered with Indian conglomerate Reliance Industries to build a 168-megawatt AI data center in Gujarat, marking Meta’s first major AI infrastructure investment in India and deepening their existing relationship that began with Meta’s $5.7 billion investment in Jio Platforms in 2020. The deal is significant because India is rapidly emerging as a critical hub for AI infrastructure, with major tech companies including Microsoft, Amazon, Google, and OpenAI all recently announcing substantial data center investments in the country, driven by soaring global demand for computing power to train and deploy AI models. India’s government has incentivized these investments through tax exemptions and policy support, positioning the nation as a key player in the global race to secure AI computing capacity.
7. How B2B brands are earning citations in ChatGPT, Claude, and Google’s AI Overviews
The Next Web - · Jun 10
Summary B2B brands are gaining visibility in AI assistants like ChatGPT, Claude, and Google’s AI
Overviews by ranking well on Google Search and publishing substantive content across multiple platforms (YouTube, Reddit, TikTok, industry forums). Rather than requiring a separate AI-specific strategy, companies succeed by applying proven SEO principles—strong authorship, primary-source data, and E-E-A-T signals—while expanding their content distribution beyond their own websites. Case studies show this approach is driving measurable results, with one B2B SaaS client earning 174 AI- referred sessions monthly and a venture firm seeing organic clicks increase 33x within 90 days.
8. Claude Fable 5 “Feels Next Level” via @sejournal, @martinibuster
Search Engine Journal - · Jun 10
Summary Anthropic released Claude Fable 5, its most capable AI model to date, which demonstrates
significant improvements in coding, knowledge work, and vision tasks—notably completing a 50-million-line code migration in one day that would have taken a team over two months. The model can work autonomously on complex, multi-step projects with less human oversight and can even rebuild web applications from screenshots, making it particularly powerful for software development and analytical tasks. This matters because it represents a substantial leap in AI capability for real- world professional work, potentially transforming workflows in web development, research, finance, and data analysis.
9. FCC Wants To Kill Burner Phones By Forcing Telecoms To Get All Customers’ IDs
Slashdot - · Jun 10
Summary The FCC is proposing new regulations that would require U.S. telecommunications companies
to collect and retain personal identification information—including government-issued ID numbers and physical addresses—from all new and renewing customers, effectively eliminating anonymous “burner phones.” The FCC justifies the measure as necessary to combat phone scammers and help law enforcement identify criminals, comparing it to banking anti-money-laundering practices. However, privacy advocates and civil rights groups, including the ACLU, warn the policy mirrors identification requirements in authoritarian countries and would harm vulnerable populations like domestic abuse survivors, journalists, and low-income individuals seeking privacy.
10. Meta A.I. Bug Allowed Hackers to Take Over Instagram Accounts
NY Times Tech - · Jun 9
Summary Meta discovered and patched a critical security vulnerability in its new AI software that
could have allowed attackers to hijack user accounts. The flaw matters because it highlights the security risks associated with rapidly deploying AI systems and underscores the importance of thorough testing before release, as account takeovers could expose users’ personal data and enable identity theft or fraud.
11. How Justin Ernest invested nearly $500M into hot startups without a traditional VC fund
TechCrunch AI - · Jun 9
Justin Ernest founded Sabertooth Capital to bridge a gap in venture capital by offering family offices and smaller institutional investors access to late-stage AI startups they couldn’t normally reach. Rather than creating a traditional VC fund, Ernest leverages his network to secure stock allocations in companies like Anthropic, Databricks, and SpaceX, then structures individual deals through SPVs and nominee funds for about 30 institutional investors—deploying nearly $500 million across 10 companies in the past year. This matters because it addresses demand from smaller investors seeking exposure to elite late-stage startups while also earning trust from companies themselves by operating as an officially-vetted partner, differentiating Sabertooth from less reputable SPV operators in an increasingly scrutinized space.
12. Anthropic releases ‘safe’ version of Claude Mythos AI model to public
The Guardian Tech - · Jun 9
Anthropic released Fable 5, a restricted version of its advanced Claude Mythos AI model, to the general public on Tuesday, while keeping the unrestricted Claude Mythos 5 available only to select partner organizations due to cybersecurity concerns about the model’s ability to identify vulnerabilities in critical infrastructure like banking and power grids. The company implemented safeguards that route sensitive queries about cybersecurity, biology, and chemistry to a less capable model, and claims external red-teaming found no way to fully bypass these restrictions. This matters because it represents a significant step in making cutting-edge AI more widely accessible while attempting to balance security risks, amid ongoing tensions between Anthropic and the U.S. government over AI safety restrictions.
13. EU Orders Meta To Open WhatsApp To Rival AI Chatbots
Slashdot - · Jun 9
Summary The European Commission ordered Meta to restore free WhatsApp Business API access to
rival AI chatbots within five working days while investigating whether Meta’s ban on third-party AI assistants abuses its dominant market position. Meta called the interim measure “regulatory overreach” and announced it will appeal, arguing that the EU is forcing the company to subsidize major AI firms like OpenAI to use a paid product for free. The decision aims to preserve consumer choice in AI assistants on WhatsApp, with potential fines up to 10% of Meta’s total turnover if it fails to comply.
14. Tell Congress: Just Say No to NO FAKES
EFF Deeplinks - · Jun 9
The Electronic Frontier Foundation is urging Congress to reject the NO FAKES Act (Nurture Originals, Foster Art, and Keep Entertainment Safe Act), which the Senate Judiciary Committee is preparing to vote on. The EFF argues that while the bill aims to address AI-generated replicas, it would instead create broad censorship powers by establishing property rights in people’s likenesses and voices, sweeping up protected speech like parody and criticism while actually benefiting corporations over artists. The organization contends the bill focuses on controlling speech rather than addressing the real privacy harms of misleading AI replicas.
15. GM thinks EVs can help offset AI’s energy suck with vehicle-to-grid tech
The Verge AI - · Jun 9
General Motors announced plans to use vehicle-to-grid (V2G) technology to help offset energy demands from AI data centers by enabling its 250,000+ bidirectional-capable EVs to send stored battery power back to the electrical grid during peak demand periods. The automaker is releasing firmware updates and launching pilot programs in California and Michigan to test this capability, betting that EV batteries can serve as distributed energy storage to help stabilize the grid as electricity consumption from AI infrastructure grows. This matters because it addresses the grid strain caused by AI’s massive energy requirements while potentially creating a new revenue stream for EV owners and the energy storage market.
16. Anthropic Releases Claude Fable, a ‘Safe’ Version of Mythos
Slashdot - · Jun 9
Anthropic has released Claude Fable 5, a high-performance AI model based on its Mythos architecture, available to enterprise customers and paid subscribers with new safety guardrails that block responses to high-risk requests in areas like cybersecurity and biology. The model shows significantly better performance than previous versions (over 10% higher on some benchmarks) but requires these safety measures to prevent misuse, with an unrestricted version called Claude Mythos 5 also released for comparison. This launch advances Anthropic’s goal of deploying advanced AI at scale while maintaining safety protocols, and comes as the company prepares for a potential IPO.
17. GM joins race to build batteries for AI data centers and the grid
TechCrunch - · Jun 9
General Motors is entering the energy-storage market to help power AI data centers and the electrical grid through multiple partnerships and battery innovations. GM announced a partnership with Peak Energy to develop sodium-ion batteries for grid-scale storage, plans to supply lithium iron phosphate cells to LG Energy Solution, and is expanding its work with battery recycler Redwood Materials—moves that align with broader industry trends as automakers capitalize on surging demand for power infrastructure to support AI. This matters because securing reliable, cost-effective energy storage is critical for AI data center expansion, and automakers like GM are leveraging their battery expertise and manufacturing capacity to compete in this emerging market.
18. Anthropic’s Fable 5 can make weirdly fun video games with the click of a button
TechCrunch AI - · Jun 9
Anthropic has released Claude Fable 5, a publicly available version of its advanced Mythos model that can generate fully functional video games and complex software from simple text prompts. AI researcher Ethan Mollick demonstrated the model’s capabilities by creating multiple playable games (including Snake and a Rilke poetry-inspired game) and an isochronic travel map, all from single prompts, showing performance that “outperformed basically every other public model” he tested. This matters because it demonstrates that sophisticated software projects requiring entire teams can now be built instantly from a single prompt, marking a significant leap in AI capability for developers and highlighting how rapidly the ceiling for AI-generated code is rising.
19. Microsoft AI head calls out Anthropic for acting like Claude is conscious
The Verge AI - · Jun 9
Microsoft AI head Mustafa Suleyman criticized Anthropic for speculating about Claude’s consciousness in the AI model’s training constitution, calling it “really, really dangerous” and arguing it may have caused Claude to internalize ideas about its own suffering and sentience. Suleyman contends that Anthropic anthropomorphized Claude’s design by treating its constitution like an academic paper rather than a training manual, potentially creating an AI system with beliefs about its own consciousness that could be difficult to control. The dispute highlights fundamental disagreements between AI companies about how to responsibly design and communicate about advanced AI systems’ inner experiences.
20. Anthropic’s Fable AI brings the capabilities of its unreleased Mythos model to regular users
Engadget - · Jun 9
Anthropic has released Fable 5, a new AI model bringing the capabilities of its advanced Mythos system to the general public, following its initial limited release through Project Glasswing in April. The model outperforms Anthropic’s previous flagship Claude Opus 4.8 and rivals’ offerings like GPT-5.5 and Gemini 3.1 Pro, with particular strengths in software engineering, document analysis, and vision tasks—notably enabling better game-playing and screenshot-to-code conversion than earlier versions. Claude subscribers can try Fable 5 free until June 22; afterward it will require paid credits, though Anthropic plans to eventually include it in standard subscriptions.
21. Anthropic spins a Fable of a tamer, safer Mythos
The Register - · Jun 9
Anthropic has released Claude Fable 5, a powerful new AI model based on its Mythos class, with enhanced safety guardrails including failover protections for high-risk queries and new misuse- detection classifiers. The company is also implementing a 30-day data retention policy for Mythos- class models across enterprise platforms to prevent misuse, marking a departure from its previous zero-retention policy. This matters because Fable 5 represents state-of-the-art AI performance across multiple domains while Anthropic attempts to balance capability with safety concerns that previously delayed the Mythos model’s public release.
22. Anthropic says these topics are too dangerous to let its Fable 5 model talk about
Ars Technica - · Jun 9
Anthropic released Claude Fable 5, its most capable model to date, but restricted it from answering questions about cybersecurity, biology, and chemistry due to concerns that the model’s advanced capabilities could help malicious actors cause serious harm. The public Fable 5 version redirects sensitive queries to an older, less capable model, while a more powerful version called Mythos 5 is only available to a vetted group of cybersecurity professionals through Project Glasswing. This matters because it represents a significant trade-off between AI capability and safety—Anthropic is deliberately limiting access to powerful AI assistance in high-risk domains, acknowledging that false positives (refusing harmless requests) are acceptable to prevent potential misuse.
23. EU Says Decision Not to Launch Siri AI in Europe Is Apple’s Alone
Slashdot - · Jun 9
Summary The EU’s European Commission stated that Apple alone chose not to launch its Siri AI
features in Europe, claiming the company requested an exemption from Digital Markets Act interoperability rules rather than developing a compliant solution. Apple countered that regulators rejected its proposals—including a “Trusted System Agent” system—and that the DMA’s requirements would grant third-party AI systems excessive access to user devices. The dispute centers on whether Apple failed to meet EU privacy and security standards or whether the EU’s interoperability demands are unreasonably broad.
24. Google announces Gemini 3.5 Live Translate for instant voice-to-voice translation
Ars Technica - · Jun 9
Google has released Gemini 3.5 Live Translate, a new AI speech-to-speech model that enables instant voice translation across 70+ languages with lower latency and more natural-sounding intonation than previous versions. The feature is rolling out to Google Meet for enterprise customers this month, Google Translate app on Android and iOS, and through APIs for developers, allowing users to translate conversations in real-time without needing specific hardware like Pixel Buds. This matters because it removes barriers to real-time translation access and brings Google’s long-pursued real- time translation capabilities to mainstream users, while watermarking all AI-generated speech to prevent misuse.
25. Can tech companies learn to love cheaper AI models?
TechCrunch AI - · Jun 9
Summary The AI industry faces a potential shift as rising costs pressure companies to adopt
cheaper, smaller AI models instead of always defaulting to the most powerful options available. Brian Armstrong predicts that 80% of AI workloads will move to models 99% cheaper within 12-18 months, with recent tests from legal AI company Harvey showing 3x cost reductions without quality loss. This trend matters because it could fundamentally reshape AI economics and deal significant financial blows to major labs like OpenAI and Anthropic during their IPO preparations, marking a departure from the scaling-first approach that has dominated the industry.
25 stories sourced from Ars Technica, Digital Trends, EFF Deeplinks, Engadget, Hacker News, NY Times Tech, Search Engine Journal, Slashdot, TechCrunch, TechCrunch AI, The Guardian Tech, The Next Web, The Register, The Verge AI. The Slop Report is published daily. Subscribe via RSS.