Nvidia launches Metropolis Blueprint as Hugging Face updates platform

Key Takeaways

Illinois Senate Democrats propose AI safety laws

Illinois Senate Democrats are introducing a new package of laws to regulate artificial intelligence and protect public safety. State Senator Bill Cunningham says the state cannot wait for federal action and must create its own standards. The legislation includes three main bills led by different senators to address specific risks. Senate Bill 315, led by Senator Mary Edly-Allen, requires large AI companies like ChatGPT and Claude to publish annual safety reports and report critical incidents within 72 hours. Senate Bill 316, led by Senator Laura Ellman, mandates that AI systems detect suicidal thoughts and connect users to crisis services like the 9-8-8 hotline. Senate Bill 317, led by Senator Rachel Ventura, requires companies to clearly tell customers when they are interacting with an AI chatbot instead of a human. These measures aim to balance innovation with safety while protecting children and consumers.

Illinois lawmakers push AI regulations before session ends

Illinois state lawmakers are rushing to pass artificial intelligence regulations before the General Assembly session ends on May 31, 2026. State Rep. Daniel Didech argues that self-regulation by tech companies has failed and led to tragic outcomes for children. The proposed laws include mandatory safety plans for large AI developers that must be reviewed by a third party. These plans would prevent AI from creating lethal weapons or urging users toward suicide. State Sen. Rachel Ventura introduced a bill requiring companies to disclose when customers are talking to an AI chatbot. Rep. Jennifer Gong-Gershowitz is pushing for AI models to be treated as products, making developers legally responsible for any harm they cause. The state hopes to follow the regulatory templates already created by New York and California.

NVIDIA launches AI tools to search and analyze video

NVIDIA has released a new system called the Metropolis Blueprint for video search and summarization to help organizations find insights in large video collections. The Video Search and Summarization system uses microservices and AI models to monitor footage and detect trends in real time. Developers can now use coding agents like Codex and OpenClaw to automatically set up and manage these video analysis tools through simple chat interfaces. The system allows users to deploy video search profiles and interact with large volumes of video data without manual configuration. NVIDIA provides a GitHub repository with skills that enable these agents to perform complex video tasks efficiently.

PG&E uses AI to stop wildfires and power outages

Pacific Gas and Electric Company launched a new Continuous Monitoring Center in San Ramon to use machine learning for wildfire prevention. The center analyzes data from 5.5 million smart meters and tens of thousands of sensors to find problems before they cause fires or outages. In 2025, the system stopped 17 potential fires and saved 12 million minutes of customer outages. The technology includes GridScope devices and SmartDetect, which use AI to monitor the electrical grid performance. Engineers recently used the system to find a wiring issue in Nevada County before it could cause an ignition. Since 2025, the center has identified 1,484 similar hazards across PG&E's network.

Hugging Face lets AI agents train models with new skills

Hugging Face has updated its platform to allow AI agents to train machine learning models and use new skills to interact with the Hub. Merve Noyan from Hugging Face explained that agents can now search for models, fine-tune them with specific datasets, and run jobs directly through the platform. The new skills feature lets agents perform tasks like building demos or exploring datasets programmatically. Users can also serve large language models locally using tools like llama.cpp for better privacy and performance. The update demonstrates how agents can automatically find the best models for tasks like reading French documents and start training processes without human intervention.

Houston faces privacy debate over new AI traffic cameras

The City of Houston approved a purchase of 10 AI-powered cameras for $350,000 to manage traffic on the Almeda corridor near NRG Stadium. Councilmember Alejandra Salinas raised concerns about mass surveillance, noting this is the second set of AI cameras the city has voted on. The cameras, made by MoboTrex, claim to use AI to track vehicles and predict violations up to two seconds in advance. Resident Joan Staunton expressed worry about having her face and license plate tracked without clear justification. University of Houston professor Kailai Wang warned that these systems could be hacked, potentially disrupting traffic signals during major events like the FIFA World Cup. The city has not yet specified which intersections will receive the cameras or when they will be installed.

SAP partners with Anthropic to build AI agents

SAP has signed a new agreement with Anthropic to help its customers build AI agents using the Claude models. This partnership aims to speed up SAP's adoption of AI agents, which has been slower than some competitors. The integration will make it easier for businesses to create and deploy intelligent agents for their operations. This move represents a significant step for SAP in the competitive AI software market.

Cybersecurity firms use AI to find code vulnerabilities

Cybersecurity researchers found that new frontier AI models are extremely effective at finding vulnerabilities in computer code. Models like Anthropic's Claude Mythos and OpenAI's GPT-5.5-Cyber scanned over 130 products and discovered 75 security issues in May 2026. These findings were much higher than usual, prompting companies to release urgent security patches. Experts warn that these powerful AI tools could become widely available to attackers within three to five months. Organizations are advised to use AI scanning on their own code and open-source supply chains to fix problems before hackers do. The industry is moving toward incorporating these AI tools directly into software development to build secure applications by design.

Economists debate if AI is distorting the economy

Economists and business leaders are divided on whether artificial intelligence is distorting the American economy. The five largest companies spent $800 billion on AI this year, but some worry this is creating economic imbalances. Norman Miller and David Ely argue that AI is transforming the economy but not distorting it, as new jobs and uses emerge constantly. Caroline Freund and Ray Major believe AI investment is crowding out spending on housing and manufacturing, creating a bubble similar to past tech booms. Kelly Cunningham sees the disruption as a necessary part of progress, even if it causes short-term pain. Bob Rauch notes that the benefits are currently flowing mostly to large companies that can afford the expensive infrastructure.

Festo introduces AI gripper for robotic handling

Festo has launched GripperAI, a software system that helps robots handle mixed products without needing to be reprogrammed for each item. The system uses a 3D camera to identify objects and automatically selects the best tool from a station of various grippers. It works with most industrial robots and can handle items ranging from small USB sticks to heavy boxes weighing 20 kilograms. If a robot misses a grip, the system automatically recalculates and tries again without stopping production. This technology helps factories like the Wurth Group in Germany manage rising product variety without expensive custom setups for every new item.

Older Americans rarely use AI despite high concern

A large portion of older Americans are not using artificial intelligence despite expressing significant worry about the technology. More than 50 percent of the Silent Generation and 39 percent of Boomers have never used AI tools. Many older adults have heard little about AI and feel little excitement about it, yet 39 percent of Boomers and 31 percent of the Silent Generation report high levels of concern. Kevin Frazier argues that policymakers should help seniors access AI tools that could improve their lives, such as autonomous vehicles for transportation. He suggests that regulators should lower barriers for services like Waymo and make rides more affordable for seniors through grants to local organizations.

Sources

• Illinois Senate Democrats highlight AI safety and privacy legislation
• State lawmakers want to rein in artificial intelligence. Here's how.
• Transform Video Into Instantly Searchable, Actionable Intelligence with AI Agents and Skills
• PG&E's new center uses machine learning to prevent wildfires
• Hugging Face: Agents Train Models with New Skills
• Privacy concerns raised as Houston approves second AI traffic camera purchase
• SAP Deepens Ties With Anthropic, Unveils AI Product Bundle
• Defender's Guide to the Frontier AI Impact on Cybersecurity: May 2026 Update
• Is artificial intelligence distorting the economy?
• Festo introduces AI-based gripper for mixed-product robotic handling
• Your grandma should be using AI. really