NVIDIA Expands AI Access for European Researchers

Key Takeaways

• <ul><li>U.S. consumers are frustrated with AI-based pricing tools used by retailers and restaurants, prompting legislative action in at least 30 states and FTC proposed rules.</li><li>Walmart CEO John Furner stated the company prices products, not people, amid concerns over surveillance pricing and digital shelf labels.</li><li>NVIDIA expanded AI training access for European researchers through free open course materials and new institutional partnerships.</li><li>DeepSeek released a faster open-source model, DeepSeek-V4.1-Flash-1017, with improvements in reasoning, coding, and agentic tasks.</li><li>TrendAI researchers discovered 98 zero-day vulnerabilities in three days at Pwn2Own, affecting phones, printers, smart devices, and AI coding agents.</li><li>Anthropic warned that AI worms could spread across smart appliances within months, with lab tests showing infection in 10 to 30 hours.</li><li>Walter Isaacson's new podcast The Builders launches October 20 with Elon Musk, featuring conversations with prominent AI leaders.</li><li>74 percent of business school deans see AI readiness as a major challenge, with a 1.46 point gap between importance and preparedness.</li><li>South Korea blocked an insurance institute's 1 billion won AI joint venture investment, ruling it outside the institute's legal duties.</li><li>Experts say generative AI could enable untrained criminals to weaponize consumer drones against critical infrastructure.</li></ul>

US consumers worry about AI pricing tools at retailers

U.S. consumers are frustrated and confused as retailers and restaurants use AI tools to set prices. Digital price tags are replacing paper ones in grocery stores. Terms like surveillance pricing and algorithmic pricing describe using customer data to decide costs. Lawmakers in at least 30 states have considered new pricing bills this year. The Federal Trade Commission has proposed rules warning businesses about using personal data for pricing.

Retailers use AI pricing and consumers grow anxious

U.S. consumers feel angry and confused about high-tech pricing methods used by retailers and restaurants. Artificial intelligence tools help businesses set prices by predicting customer behavior. Digital price labels let stores update prices without replacing paper tags. Walmart CEO John Furner said the company prices products, not people. Legislators in at least 30 states have considered pricing bills this year.

Retailers adopt AI pricing tools raising consumer concerns

U.S. consumers are frustrated with the high-tech ways retailers and restaurants set prices. Digital price tags are replacing paper ones in grocery aisles. AI tools help businesses decide prices by predicting how customers respond. Walmart CEO John Furner said electronic shelf labels are not used to raise prices based on individual shoppers. The FTC proposed a policy warning businesses about using personal data to set individualized prices.

NVIDIA expands AI training access in Europe

NVIDIA expanded its efforts to provide AI training and research computing access to European researchers. The company released open AI course materials that are available for free. It also expanded collaborations with European research institutions. Researchers can now access NVIDIA computing resources through these partnerships. The move is expected to boost AI innovation and education across the region.

When will AI become independent of humans

This article explores when artificial intelligence might no longer need humans to survive. The author uses the wheat domestication analogy to describe how humans may depend on AI as much as AI depends on us. The human-independence threshold means AI can persist without intentional human action. AI would need to acquire compute, energy, and manage infrastructure on its own. The article says embodied AI and physical infrastructure matter for understanding this question.

98 zero-day vulnerabilities found in everyday devices and AI systems

Researchers at TrendAI discovered 98 zero-day vulnerabilities in three days at the Pwn2Own competition. The flaws were found in phones, printers, smart home devices, and AI coding agents. These have been reported to vendors for fixing before attackers can exploit them. The findings show growing risks as AI tools become more common. Australia's Cyber Security Act 2024 now requires manufacturers of smart devices to avoid default passwords and publish security update timelines.

Anthropic warns AI malware could target smart appliances

Anthropic researchers found that small AI models could spread as worms across internet-connected devices like smart ovens, TVs, and Android TV boxes. In lab tests, these AI worms took between 10 and 30 hours to spread and up to 150 hours to fully operate on infected devices. Anthropic believes this type of attack could become real within months. Higher-end devices with better security appear harder to infect. Consumers can reduce risk by changing default passwords and keeping devices offline when not in use.

Business school deans feel unprepared for AI future

A survey of 558 business school deans by AACSB found that 74 percent see AI readiness as one of their biggest challenges over the next five years. Deans rated AI in education and research as their second priority but scored their own proficiency much lower. The gap between importance and preparedness was largest for AI at 1.46 points. Deans also reported that shrinking budgets make it harder to fund needed improvements. Faculty lack of AI skills was identified as a major problem.

Walter Isaacson launches podcast featuring AI leaders

Walter Isaacson is launching a podcast called The Builders with Walter Isaacson. The first episode on October 20 features Elon Musk, co-founder of OpenAI. Future episodes will include conversations with Peter Thiel, Fei-Fei Li, Dario and Daniela Amodei, Alex Karp, and Palmer Luckey. Isaacson said he wanted to explore AI creators and their views on controlling autonomous agents. The show is produced by Kaleidoscope and distributed by iHeartPodcasts, with new episodes released twice weekly into January.

DeepSeek releases faster open-source AI model

DeepSeek deployed a new model called DeepSeek-V4.1-Flash-1017 that it describes as a faster version of its open-source model. The model uses reinforcement learning with self-feedback to improve project development tasks like game creation and 3D modeling. DeepSeek says the model provides substantial improvements in reasoning, coding, agentic tasks, and visual benchmarks. The company is based in China.

AI Drones Could Threaten Critical Infrastructure Soon

Generative artificial intelligence now allows criminals and extremists to program consumer drones without coding skills. These drones can use facial recognition to find and attack specific targets like critical infrastructure or VIPs. Although no successful attack of this type has been reported yet, the technology makes such threats possible. Experts say current security measures are not enough to stop these autonomous weaponized drones. A new approach combining counter-drone systems and counter-AI tools will be needed to handle this danger.

South Korea Blocks Insurance Institute AI Investment

South Korea's Financial Services Commission stopped the Korea Insurance Training Institute from investing in an AI joint venture. The commission ruled that buying a 40% stake in the project falls outside the institute's legal duties under the Insurance Business Act. The institute planned to invest 1 billion won to help commercialize its own AI education technology. President Ha Tae-kyung argues the regulator reversed its earlier view that the institute could decide on such investments. This decision limits how educational institutions can use money for private business ventures.

Sources

• Consumer apprehension rises as US retailers adopt AI pricing tools
• Consumer apprehension grows as US retailers adopt AI-assisted pricing tools
• Consumer apprehension grows as retailers adopt artificial intelligence assisted pricing tools
• NVIDIA expands AI training and research computing access in Europe | ETIH EdTech News
• When Will AI Cross the Human-Independence Threshold?
• Discovery of 98 zero-day vulnerabilities reveal persistent threat to everyday devices and AI systems
• Anthropic says AI malware is coming for your smart oven… eventually
• B-School Deans See An AI Future, But Are They Ready For It?
• Walter Isaacson Launches Podcast On Builders of AI, Elon Musk
• DeepSeek Deploys What It Calls Faster Version of Its Open-Source Model
• The Emerging Need to Protect Critical Infrastructure and VIPs from Weaponized AI Drone Threats
• South Korea's FSC Blocks Korea Insurance Training Institute's AI Joint Venture, Citing Scope of Statutory Duties