Digimasters Shorts - Swedish PM’s AI Risks Spark Backlash, DeepMind’s Genie 3 Revolution, Perplexity AI Called Out by Cloudflare, OpenAI’s ChatGPT Mental Health Push, Tesla’s JB Straubel Reinvents Battery Power

Welcome to Digimasters Shorts, where we serve up the latest and most intriguing updates from the digital world in quick, digestible episodes. Join hosts Adam Nagus and Carly Wilson as they explore the intersection of AI, technology, and society—covering everything from politicians using AI for decision-making to cutting-edge innovations like DeepMind's Genie 3. Stay informed on how major players like Google DeepMind, OpenAI, and Cloudflare are shaping the future of AI, while also diving into real-world applications such as sustainable energy solutions powered by recycled EV batteries. Whether you're a tech enthusiast, industry insider, or just curious about the digital frontier, Digimasters Shorts delivers the essential scoops you need in under 15 minutes. Tune in for your daily dose of digital intelligence!

Transcript

Adam N2: 0:05

Welcome to Digimasters Shorts, we are your hosts Adam Nagus

Carly W: 0:09

and Carly Wilson delivering the latest scoop from the digital realm. Swedish Prime Minister Ulf Kristersson has revealed he frequently uses AI tools like Chat G.P.T and LeChat for a second opinion in government matters. Kristersson's center-right coalition colleagues also incorporate AI in their daily tasks. In an interview with Dagens industri, he explained it helps to consider different perspectives. However, tech experts have expressed concerns about politicians relying on AI for decision-making. The newspaper Aftonbladet accused Kristersson of succumbing to"the oligarchs' AI psychosis." Computer scientist Simone Fischer-Hübner warned against using AI with sensitive information. Kristersson's spokesperson clarified that no security-sensitive data is input into AI tools, describing their use as more of a"ballpark" reference. Meanwhile, AI ethics professor Virginia Dignum stated AI cannot provide meaningful political insight, as it merely reflects its creators' views. She cautioned that overreliance on AI risks growing overconfidence in its capabilities. Dignum emphasized that voters elected a human leader, not an AI system.

Adam N2: 1:16

Google DeepMind has unveiled Genie 3, the latest iteration of its generative AI world model, just seven months after releasing Genie 2. Genie 3 can create interactive, changeable 3D environments from simple prompts or images, allowing users to alter objects, weather, and characters in real time. While this technology offers potential in game development, many in the gaming industry remain skeptical of its practical benefits. DeepMind views Genie 3 primarily as a research tool to help train AI agents in complex, dynamic worlds, moving closer to artificial general intelligence. Unlike its predecessor, Genie 3 boasts higher visual fidelity, real-time navigation at 720p and 24 frames per second, and significantly improved memory lasting multiple minutes. This extended retention overcomes Genie 2’s 10-second memory limitation, enabling more coherent and consistent simulations. However, Genie 3 still struggles with AI hallucinations, inaccuracies in human movement, and nonsensical text generation. Current AI agents can only navigate these worlds but lack the reasoning to modify the simulation itself, with DeepMind exploring multi-agent interactions for future versions. The heavy computational cost means Genie 3 remains inaccessible to the public and is limited to select researchers for now. DeepMind plans to gradually expand access as it continues refining this promising world model technology. Search engine provider Perplexity AI has been accused by Cloudflare of violating anti-scraping rules while crawling websites. Cloudflare's C.E.O Matthew Prince likened Perplexity A.I's behavior to that of"North Korean hackers" due to their invasive web crawling practices. An investigation revealed that Perplexity AI repeatedly modified its bots to evade measures meant to block unauthorized data scraping. In response, Cloudflare removed Perplexity AI from its list of verified bots and updated systems to block its stealth crawling activities. Cloudflare verified these findings by setting up test sites hidden from search engines, which Perplexity AI still accessed using disguised user agents and rotating IP addresses. This activity spanned tens of thousands of domains and millions of requests per day. The crackdown threatens Perplexity A.I's ability to index content and underscores growing tensions over AI companies harvesting data without permission. Media companies have filed lawsuits against Perplexity AI and others for copyright infringement concerning AI training data. Cloudflare expects Perplexity AI to continue adapting its crawling tactics, unlike companies like Open A.I which respect anti-scraping rules. The incident highlights the ongoing conflict between A.I's data demands and the rights of content owners.

Carly W: 4:00

Open A.I has introduced new features to its Chat G.P.T to promote healthier user interactions and support well-being. The AI will now send gentle reminders during long sessions, encouraging users to take breaks. Chat G.P.T is also improving its ability to detect signs of mental or emotional distress and will guide users to appropriate resources without worsening the situation. For sensitive questions involving high-stakes personal decisions, Chat G.P.T will avoid giving direct answers and instead help users weigh pros and cons. Open A.I is consulting with mental and physical health experts to better handle vulnerable users during conversations. These enhancements are part of an ongoing effort and may roll out gradually. The updates aim to help users stay in control of their time and manage challenges with guidance, not decisions. Open A.I emphasizes that Chat G.P.T’s role is to support, not replace professional advice. The improvements come amid anticipation of the potential G.P.T-5 launch in the coming weeks. This initiative reflects growing use of AI chatbots for emotional support across various domains. As demand for energy soars with the rise of AI, JB Straubel, Tesla's co-founder, is turning to a new solution: repurposing old electric vehicle batteries. His company, Redwood Materials, originally focused on recycling EV batteries, found most had remaining capacity suitable for second use. Instead of recycling them immediately, Redwood now uses these batteries for microgrid energy storage, targeting data centers. In partnership with Crusoe, known for its AI data center in Texas, Redwood deployed North America’s largest microgrid, delivering 12 megawatts of solar power and 63 megawatt-hours of battery capacity. This project supports 24/7 renewable energy at a cost-effective price, addressing long-standing battery cost challenges. With AI expected to increase data center power needs by 165% by 2030, the demand for scalable energy storage is critical. Redwood currently holds over a gigawatt-hour of reusable batteries and plans to scale up projects to over 100 megawatts. The company aims to compete in the growing energy storage market alongside giants like Tesla. Experts agree there is ample room for newcomers, especially for smaller, cost-sensitive projects using second-life batteries. As AI fuels energy demand, Redwood Materials’ innovative approach offers a promising path to sustainable power solutions.

Don: 6:37

Thank you for listening to today's AI and Tech News podcast summary... Please do leave us a comment and for additional feedback, please email us at podcast@digimasters.co.uk You can now follow us on Instagram and Threads by searching for@DigimastersShorts or Search for Digimasters on Linkedin. Be sure to tune in tomorrow and don't forget to follow or subscribe!