Congresswoman Yvette Clarke is pushing several federal AI bills, including the REAL Political Ads Act, which would require disclosure when generative AI appears in political ads. She also co-sponsored measures targeting algorithmic bias, deepfakes, and intellectual property theft. Clarke led 44 members of Congress in a letter opposing any moratorium on state AI laws embedded in the FY26 NDAA, signaling strong bipartisan interest in keeping state-level regulation alive.
OpenAI and Microsoft are defending their use of copyrighted news content, arguing it qualifies as fair use. The News/Media Alliance warns this practice threatens publisher revenue, especially as ChatGPT and similar tools deliver news summaries that keep readers from visiting original sites. The New York Times filed suit against both companies in 2023 over training data, and the debate shows no sign of cooling.
Google DeepMind CEO Demis Hassabis thinks schools should lean into AI for rote learning, letting personalized tools handle memorization at each student's pace. He envisions classrooms where a few world-class teachers deliver core lectures while others guide project-based work. His vision borrows from Montessori and Oxbridge tutorial models, aiming to sharpen creativity and judgment in an AI-driven era.
Data centres consumed 485 terawatt hours of electricity in 2024, roughly 1.5 percent of global supply, and they produce enormous waste heat. Engineers in Denmark are now capturing heat from the Bitten supercomputer at the University of Southern Denmark and routing it into Sonderborg's district heating network, warming the town's 75,000 residents. It is a practical answer to the energy footprint of AI infrastructure.
Trevor Noah told audiences at Boston's Museum of Science on Oct. 19 that AI often looks deeper than it really is and that in-person community offers the strongest defense against its downsides. He expects new jobs to appear but stresses the painful gap between roles disappearing and those emerging. Separately, the Australian Information Security Association warned firms not to skip cyber basics in pursuit of AI tools, noting the Australian Cyber Security Centre handled over 1,500 incidents in 2025-26, up more than 20 percent year over year.
Key Takeaways
- Yvette Clarke reintroduced the REAL Political Ads Act and co-sponsored four other AI bills targeting deepfakes, bias, and IP protection in Congress.
- Clarke led 44 members of Congress in opposing a moratorium on state AI laws in the FY26 NDAA.
- OpenAI and Microsoft claim fair use for training on copyrighted news content, while the New York Times sued them in 2023 over the practice.
- Harvard Business Review research on 515 companies found teams that redesigned work around AI identified 44 percent more use cases and nearly doubled revenue.
- Demis Hassabis proposes schools use AI for rote learning so students can focus on creativity, projects, and judgment.
- Vellum removed its AI automation feature entirely in version 4.2.1 after strong backlash from authors.
- The Australian Cyber Security Centre responded to over 1,500 incidents in 2025-26, a 20 percent increase from the prior year.
- AI data centres used 485 terawatt hours of electricity in 2024, about 1.5 percent of the world's total energy supply.
- Danish engineers now use waste heat from the Bitten supercomputer to warm 75,000 residents in Sonderborg.
- Trevor Noah warns the gap between disappearing and emerging AI-era jobs is a serious concern.
Rep. Yvette Clarke pushes federal AI legislation
Congresswoman Yvette Clarke has introduced multiple AI bills in Congress. She reintroduced the REAL Political Ads Act requiring disclosure when generative AI is used in political ads. She also co-sponsored the AI Civil Rights Act, the Algorithmic Accountability Act, the CREATOR Act, and the DEEPFAKES Accountability Act. These bills target algorithmic bias, deepfakes, and intellectual property protection. Clarke also led 44 members in a letter opposing any moratorium on state AI laws in the FY26 NDAA.
AI companies defend stealing news as fair use
The News/Media Alliance warns that AI companies like OpenAI and Microsoft are using copyrighted news content without permission. The New York Times sued OpenAI and Microsoft in 2023 for training AI models on their content without permission. AI chatbots like ChatGPT provide summaries or copies of news that replace visiting news sites, hurting publisher revenue. OpenAI and Microsoft argue their use qualifies as fair use, but the article argues this interpretation threatens the news industry and violates copyright law.
Vellum removes AI automation after author backlash
Vellum, a Mac-based book formatting tool, reversed its decision to add AI automation features after strong criticism from authors. The feature allowed users to connect external AI assistants to handle repetitive publishing tasks. The three-person development team admitted they underestimated how strongly users objected to any AI integration. In version 4.2.1, the Automation feature was entirely removed from the software.
Judgment matters more than cheap AI tools
An article argues that while every company now has AI, the real advantage comes from human judgment. Sid Pednekar says companies that redesign work around AI rather than just automating old steps perform better. Research from Harvard Business Review involving 515 companies found teams that redesigned work around AI identified 44 percent more use cases and generated nearly twice as much revenue. Pednekar also notes that 71 percent of organizations are modernizing core systems for AI, but data quality remains a key challenge.
Demis Hassabis says schools should use AI for rote learning
Google DeepMind CEO Demis Hassabis proposes restructuring education for an AI-driven future. He says schools should focus on projects, creativity, people skills, and judgment instead of memorization. Rote learning should be handled by personalized AI tools at each student's own pace. Hassabis also suggests a few world-class teachers should deliver core lectures while other teachers focus on project-based classroom activities. His vision draws on Montessori-style classrooms and Oxbridge tutorial models.
Trevor Noah shares insights on AI and jobs
Trevor Noah spoke about AI at Boston's Museum of Science on Oct. 19. He said AI often seems deep but only works on the surface. He believes in-person community is the best defense against AI downsides. He is optimistic that new jobs will emerge but worries about the gap between old and new jobs.
AISA warns firms not to skip cyber basics for AI
The Australian Information Security Association warned organizations not to prioritize AI security tools over basic measures. It spoke out after the Australian Signals Directorate reported a sharp rise in cyber activity. The Australian Cyber Security Centre responded to over 1,500 incidents in 2025-26, up more than 20 percent from the previous year. Experts said fixing unpatched vulnerabilities and poor system visibility matters more than buying new AI products.
How to spot AI-generated music
AI music is becoming common on streaming platforms, but there are ways to detect it. AI vocals often sound slurred, warbled or monotonous and lack natural human imperfections. AI instrumentals usually miss dynamic shifts, key changes and non-Western instruments. AI lyrics tend to repeat ideas, avoid metaphor and lack a clear theme. Checking an artist's online presence can also help reveal fake or content-farm accounts.
AI data centre waste heat could warm homes
AI data centres use huge amounts of energy and produce a lot of waste heat. Engineers in Denmark are using liquid cooling technology from Danfoss to capture heat from the Bitten supercomputer at the University of Southern Denmark. The recovered heat goes into Sonderborg's district heating network to help warm the town's 75,000 residents. The International Energy Agency reported data centres used 485 terawatt hours of electricity in 2024, about 1.5 percent of the world's total energy supply.
Sources
- Rep. Yvette Clarke Stands on the Frontlines of AI Legislation
- Opinion/Guest column: When AI confuses theft of news with 'fair use'
- MacBook Book Formatting Tool, Vellum, Reverses Course On AI Automation After Author Backlash
- AI Is Cheap. Judgment Is the Advantage
- Schools Should Eventually Be About Projects & Creativity, Personalized AI Will Take Care Of Rote Learning: Demis Hassabis
- Trevor Noah on AI
- AISA warns firms not to put AI before cyber basics
- AI music can be difficult to detect, but there are some telltale signs
- As AI data centres guzzle energy, can we harness their wasted heat?
Comments
Please log in to post a comment.