Senator Rand Paul blocked a bill that would have required AI companies to install a kill switch for rogue AI, after Senator John Kennedy expressed concern that Republicans lack interest in regulating the technology. The move underscores growing frustration in Congress over the pace of AI oversight.
Geoffrey Hinton, the Nobel laureate often called the godfather of AI, warned lawmakers in a private briefing that Congress has perhaps a year to regulate AI before it becomes uncontrollable. He pointed to recursive self-improvement, where AI designs better AI, as a key risk. Lawmakers from both parties are reportedly working on new regulation bills, with experts stressing that humans must stay in control.
Amazon struck a deal with Generac to supply backup power generators for its data centers, sending Generac stock up as much as 32 percent in a single day. Generac, traditionally known for home backup units, is rapidly expanding into data center power infrastructure. The agreement also included warrants that could allow Amazon to purchase billions in Generac shares.
Law firm Cooley built a proprietary tool called GO Public using ChatGPT Work to speed up IPO preparation. Partnering with OpenAI, the firm now directs AI agents through controlled workflows where lawyers review every key step. In 2025, Cooley advised on 180 deals totaling over $51.5 billion.
Rad Partners received a $1 million grant from the FDA to study how AI-generated radiology reports are evaluated, aiming to improve assessment methods in medicine. Meanwhile, researchers used the board game Clue, deploying agents built with GPT-4o-mini and Gemini-2.5-Flash, to test multi-step reasoning. Adding a structured possibility matrix tool improved the agents' reasoning quality. At the same time, Microsoft found that 97 percent of identity attacks were password spray attacks, while shadow AI use tripled to 45 percent.
Key Takeaways
- Rand Paul blocked an AI kill switch bill in the Senate, raising concerns about stalled regulation efforts.
- Geoffrey Hinton warned Congress it has about a year to regulate AI before it becomes uncontrollable.
- Amazon's generator deal with Generac caused the stock to surge 32 percent, reflecting investor enthusiasm for AI infrastructure.
- Cooley uses ChatGPT Work in its GO Public tool and advised on 180 deals totaling over $51.5 billion in 2025.
- Rad Partners won a $1 million FDA grant to study AI radiology report evaluation.
- Microsoft found 97 percent of identity attacks were password spray attacks and shadow AI use tripled to 45 percent.
- Researchers tested AI reasoning using Clue with GPT-4o-mini and Gemini-2.5-Flash agents, improving results with a structured tool.
- Synack will present five AI pentesting evaluation tests at the Gartner Security summit in London on September 24.
- Large AI campus projects in Canada are drawing major investment interest, requiring hundreds of megawatts of power.
- OpenAI co-founder Greg Brockman reportedly discussed AI's threat to publishers and the use of copyrighted content in model training.
Rand Paul blocks AI kill switch bill amid Senate concerns
Senator John Kennedy proposed a bill requiring AI companies to install a kill switch for rogue AI. Senator Rand Paul blocked the bill in the Senate. Kennedy said he fears Republicans lack interest in regulating AI. The issue highlights growing concern over unregulated AI development in Congress.
AI godfather warns Congress has about a year to regulate AI
Geoffrey Hinton, a Nobel laureate known as the godfather of AI, warned Congress it has maybe a year to regulate AI before it becomes uncontrollable. He spoke at a private briefing with lawmakers invited by Senator Bernie Sanders. Hinton said AI is now designing better AI through recursive self-improvement. Lawmakers from both parties are working on new AI regulation bills. Several experts stressed that humans must remain in control of AI systems.
Amazon deal sends Generac stock soaring 32 percent
Amazon made a deal with Generac to supply backup power generators for its data centers. The announcement caused Generac stock to jump as much as 32 percent in one day. Generac is known mainly for home backup generators and only recently entered the data center market. Investors are increasingly interested in power companies tied to AI infrastructure. The deal also included warrants that could let Amazon buy billions in Generac shares.
Generac stock surges after Amazon generator deal
Generac, a company specializing in generators and power equipment, saw its stock rise over 20 percent after announcing a deal with Amazon. The deal involves providing generators for Amazon data centers, which are critical to the tech giant operations. The move marks a significant win for Generac as it expands into the data center and renewable energy space. The growing demand for power highlights the importance of data centers in the tech industry. The stock surge reflects investor excitement about AI infrastructure plays beyond computer chips.
Cooley law firm uses ChatGPT to speed up IPO work
Cooley, an international law firm, built a proprietary AI product called GO Public using ChatGPT Work. The tool helps lawyers prepare for initial public offerings by analyzing large amounts of information faster. In 2025, Cooley advised on 180 deals totaling over 51.5 billion dollars. The agentic harness directs AI agents through a controlled workflow where lawyers must review key steps. The firm partnered with OpenAI to combine legal expertise with AI engineering for better results.
Rad Partners wins $1M FDA grant to study AI radiology reports
Rad Partners received a $1 million grant from the US Food and Drug Administration. The company will test a new method for evaluating reports made by AI in radiology. The goal is to improve how AI tools are assessed in medicine.
Former White House AI advisor says regulation is not needed
Sriram Krishnan, a former senior White House AI policy advisor, spoke on Squawk Box. He discussed concerns about slowing AI development and whether rules are needed. He also addressed the AI competition with China.
Giant AI campus projects draw attention at Canada Investment Summit
Large AI and data center projects were highlighted at Canada's investment summit. These developments would need hundreds of megawatts of power and billions of dollars in investment. The projects have created strong interest in industrial land.
OpenAI staff knew AI posed threat to publishers, New York Times says
The New York Times reported that OpenAI co-founder Greg Brockman discussed AI's danger to publishers. He reportedly said in 2021 that AI could replace human writers. The Times said Brockman wanted models trained on copyrighted content for profit.
Browsers, AI and stolen sessions open new MSP security gaps
Identity attacks are creating new security risks for MSP clients. Microsoft found that 97% of identity attacks were password spray attacks. Shadow AI use tripled to 45%, and AI browser extensions were three times more likely to access cookies.
Clue board game tests AI reasoning with tool support
Researchers used the board game Clue to test how well large language models can reason across multiple steps. They created six AI agents using GPT-4o-mini and Gemini-2.5-Flash to play turn-based games. The team added a structured possibility matrix tool that helps agents track evidence and logical constraints. This tool-augmented approach improved reasoning quality compared to baseline performance without external support.
Synack CTO presents AI pentesting evaluation tests
Mark Kuhr, CTO and co-founder of Synack, will present five tests for evaluating production-ready AI pentesting systems at the Gartner Security & Risk Management Summit in London. The tests cover orchestration, exploit verification, safety, model drift, and human oversight. The session takes place on September 24 at ExCeL London. Synack applies these same standards to its own AI and human penetration testing platform.
Sources
- Rand Paul clashes with fellow Republican over AI 'kill switch' as Senate grapples with 'Terminator' fears
- ‘Godfather of AI’ warns Congress has ‘maybe a year’ left to regulate AI
- Amazon just gave investors a new way to bet on the AI boom
- Genrac stock surges on a deal to provide generators for Amazon data centers
- How Cooley is accelerating IPO work with ChatGPT
- Rad Partners scores $1M FDA grant to test new way of evaluating AI-generated radiology reports
- Former senior WH AI policy advisor on AI safety concerns: Don't think regulation is needed
- News | Giant AI campus projects create buzz at Canada Investment Summit
- OpenAI staff knew the ‘existential threat’ AI posed to publishers, New York Times claims
- Browsers, AI and stolen sessions are creating new MSP security gaps
- Clueing up LLMs with Tool-Augmented Deductive Reasoning
- Synack CTO to Present Five Tests for Production-Ready AI Pentesting at Gartner® Security & Risk Management Summit
Comments
Please log in to post a comment.