AI News, 2 April 2026
The 18 most significant AI stories of the day, selected and summarised automatically by AI. Each links to the original reporting.
- Business
OpenAI acquires TBPN, the buzzy founder-led business talk show
auto_awesomeAI SummaryOpenAI has acquired TBPN, a popular founder-focused business podcast, positioning it under political strategist Chris Lehane's oversight while allowing editorial independence. This move signals OpenAI's expanding interest in media and content platforms beyond core AI products, potentially leveraging TBPN's influential audience to shape narratives around AI development and business leadership.
- Products
PSA: Anyone with a link can view your Granola notes by default
auto_awesomeAI SummaryGranola's AI note-taking app shares notes with anyone holding a link despite claiming privacy-first design, and uses notes for AI training without explicit opt-in. This highlights growing tensions between AI product convenience and user data protection, signaling potential regulatory and trust issues for the emerging AI productivity tool market.
- Research
New Rowhammer attacks give complete control of machines running Nvidia GPUs
auto_awesomeAI SummaryResearchers have discovered critical Rowhammer-based attacks (GDDRHammer, GeForce, and GPUBreach) that exploit GPU memory to gain complete control of machines with Nvidia GPUs. These vulnerabilities represent a significant security risk for AI infrastructure and data centers relying on GPU computing, potentially allowing attackers to bypass security measures and access sensitive information.
- Products
Microsoft takes on AI rivals with three new foundational models
auto_awesomeAI SummaryMicrosoft's newly formed AI group has released three foundational models capable of transcribing voice to text, generating audio, and creating images. This move demonstrates Microsoft's effort to compete directly with other major AI players by offering diverse multimodal capabilities across different data types.
- Products
KiloClaw targets shadow AI with autonomous agent governance
auto_awesomeAI SummaryKiloClaw addresses the growing problem of shadow AI by providing enterprises with governance tools to manage autonomous agents deployed by employees without official approval. As companies formalized LLM security, workers began bypassing procurement to deploy their own agents, creating compliance and security risks that KiloClaw aims to solve.
- Research
Gemma 4: Byte for byte, the most capable open models
auto_awesomeAI SummaryDeepMind has introduced Gemma 4, an advanced open-source model designed specifically for complex reasoning and autonomous agent applications. This release represents a significant milestone in making state-of-the-art AI capabilities accessible to the broader community, potentially accelerating innovation in reasoning-heavy AI tasks.
- Products
Google now lets you direct avatars through prompts in its Vids app
auto_awesomeAI SummaryGoogle is enhancing its Vids app with prompt-based avatar customization, enabling users to generate and direct AI avatars through natural language instructions. This advancement democratizes video production by reducing technical barriers and showcasing practical applications of generative AI in content creation tools.
- Business
5 best practices to secure AI systems
auto_awesomeAI SummaryAs AI systems become critical to business operations, companies must adopt multi-layered security strategies to address vulnerabilities unique to this technology. Traditional security frameworks are insufficient for protecting AI systems, making robust defense practices essential as the industry matures.
- Business
Elon Musk is about to be a very busy boy!
auto_awesomeAI SummaryReports suggest SpaceX could pursue an IPO filing in June, potentially accelerating Musk's timeline for taking the company public. This move could reshape the commercial space industry and provide significant capital for AI-driven space technologies and autonomous systems development.
- Business
OpenAI just bought TBPN
auto_awesomeAI SummaryOpenAI has purchased TBPN, a daily live talk show that interviews AI executives and tech leaders. The acquisition gives OpenAI direct control of a platform that has hosted prominent figures from Meta, Microsoft, and other major tech companies, potentially amplifying its voice in industry conversations.
- Policy
China’s Five-Year Plan details the targets for AI deployment
auto_awesomeAI SummaryChina has officially outlined its AI priorities in its new 15-year economic plan, positioning artificial intelligence alongside quantum computing and biotechnology as critical technological frontiers. The plan details specific deployment targets and strategic directions for AI development, signaling the country's commitment to maintaining competitive advantage in this transformative technology sector.
- Business
The Download: plastic’s problem with fuel prices, and SpaceX’s blockbuster IPO
auto_awesomeAI SummarySoaring fossil fuel prices due to geopolitical tensions are creating economic ripple effects that could dramatically increase plastic manufacturing costs. This supply chain disruption highlights how energy costs directly impact material production, affecting industries reliant on AI-driven optimization and forecasting.
Read the original at MIT Technology ReviewKey takeaways & FAQ
- Business
It’s not easy to get depression-detecting AI through the FDA
auto_awesomeAI SummaryKintsugi, a seven-year-old startup developing AI to detect depression and anxiety from speech patterns, is closing due to inability to secure FDA clearance in time. The company is open-sourcing its technology, highlighting the significant regulatory hurdles AI health tools face and the challenges of commercializing medical AI applications.
- Business
Fuel prices are soaring. Plastic could be next.
auto_awesomeAI SummaryGeopolitical disruptions in the Middle East are driving fossil fuel price surges that will cascade into plastic production costs. This economic shock has implications for AI hardware manufacturing, supply chains, and the material costs of computing infrastructure that powers AI systems.
Read the original at MIT Technology ReviewKey takeaways & FAQ
- Business
Autonomous AI systems depend on data governance
auto_awesomeAI SummaryAs AI systems become increasingly autonomous, data governance is emerging as critical to AI safety—rivaling traditional model training and monitoring in importance. Poor data quality, fragmentation, and lack of oversight can make autonomous AI systems behave unpredictably, highlighting why organizations must prioritize robust data management practices alongside model safety measures.
- Business
Experian uncovers fraud paradox in financial services’ AI adoption
auto_awesomeAI SummaryExperian's 2026 Future of Fraud Forecast reveals a critical paradox: the same AI technology that financial institutions deploy to prevent fraud is being exploited by criminals to commit it. This dual-use challenge highlights the urgent need for the industry to balance AI innovation with robust security measures as fraud losses continue mounting.
- Business
Microsoft’s new ‘superintelligence’ game plan is all about business
auto_awesomeAI SummaryMustafa Suleyman, Microsoft's inaugural AI CEO, has shifted his focus toward developing superintelligence following a March restructuring that reallocated some of his previous responsibilities. This strategic move signals Microsoft's commitment to advancing AI capabilities beyond current models, reflecting the tech industry's broader race toward more powerful AI systems.
- Products
Google Home’s latest update makes Gemini better at understanding your commands
auto_awesomeAI SummaryGoogle has updated its Home app to make Gemini AI better at interpreting natural language commands for smart home devices. Users can now describe desired settings conversationally—like requesting "the color of the ocean" for lighting—and Gemini will understand and execute these requests more reliably. This advancement demonstrates how AI assistants are becoming more intuitive and accessible for everyday consumer applications.