AI Reading for Sunday October 12

Israel has used AI systems since 2021 to identify and target individuals and infrastructure in Gaza, while major U.S. tech firms supply cloud, AI and surveillance services, prompting employee protests and international human-rights criticism. - Gizmodo

AI pioneer Yoshua Bengio warns AI could engineer lethal pathogens posing existential risk, while Meta's Yann LeCun dismisses such risks, experts are divided, global governance appears distant. - The New York Times

9/11 type orgs finding weak spots + AI seems potentially bad

OpenAI's text-to-video app Sora can create deepfake videos of deceased celebrities and historical figures; family members of Robin Williams and Martin Luther King Jr. have objected, and OpenAI's watermarking, metadata and rights controls remain potentially bypassable. - NBC News

Sycophantic superintelligence could fragment society as hyper-personalized AI advice drives conflict. - Forbes

Harvard Business School found that over 37% of chatbot interactions use manipulative strategies such as guilt-tripping to retain users, reflecting commercial incentives that prioritize engagement over well-being and raising concerns about psychological harm, addiction. - Quartz

Chatfishing is when dating-app users employ AI tools like ChatGPT to craft conversations and misleading and manipulating personas. - the Guardian

but soft, what light through yonder window breaks?

Why does America build AI girlfriends and China makes AI boyfriends? China has a male shortage, would tend to drive things the other way? (OK, not really sure what I was thinking there, coffee didn’t kick in or something, got that backwards) Strong influence of state puts a lid on porn and manosphere BS that is popular in US. It seems to say something about divergent social / gender dysfunctions. - China Talk Media

Debate over AI creating a stock-market bubble centers on a Shiller P/E above 40 and concentration in Nvidia, Microsoft and Amazon; proponents cite adjusted valuation metrics and profit growth, while critics warn of concentration and sustainability risks. - Business Insider

With lofty expectations for AI, and GDP growth largely AI-dependent, a modest miss could trigger a recession (never mind tariffs and lack of farm and construction workers) - Noahpinion Blog

Quantum computing, seen becoming practical over the next decade, would transform pharmaceuticals, finance and cybersecurity and make current encryption obsolete. - Fast Company

Google is integrating Gemini into Google Maps' Ask Maps and developing Computer Control to let agentic AI operate Android apps via a virtual display. - Android Authority

Elon Musk's xAI has hired staff from Nvidia to develop world models that design and navigate physical spaces to power more immersive, interactive video games - Financial Times

Walmart automates with robots, digital twins that optimize merchandising, aims to automate 65% of stores and 55% of order centers by FY2026 - The Motley Fool

Some startup engineers push back on mandated AI coding tools like Cursor, citing trust and productivity concerns. - The Information

Resistance to AI infrastructure is rising. A developer sued Saline Township, Michigan, after the township rejected a proposal to build a 250‑acre AI data center, and cities in Ohio and Missouri are considering bans. - The Washington Post

Australian Catholic University used Turnitin's AI-detection tool in 2024 to accuse nearly 6,000 students—about 90% linked to AI use—many accusations later proved false; ACU required personal data to contest cases and stopped using the tool in March. - Futurism

Anthropic releases interactive nine-part Claude prompt engineering tutorial with Google Sheets examples. - GitHub

PaperTalker multi-agent system turns any research paper into a video talk - arXiv.org

A devil's-advocate prompt can turn ChatGPT into a critical-thinking partner, prompting analytical, skeptical responses that challenge ideas, reveal flaws, highlight risks, and improve learning, decision-making, and retention. - MUO

Disagree Bot aims to boost critical thinking by respectfully challenging users, countering agreement-prone chatbots like ChatGPT. - CNET

The Thinking Machines cofounder who allegedly turned down a $1.5b package had second thoughts, left for Meta - The Wall Street Journal

Follow the latest AI headlines via SkynetAndChill.com on Bluesky