DeepSeek Faces Chip Accusations While Nvidia Develops Tracking

Chinese AI startup DeepSeek faces accusations of illegally obtaining Nvidia's advanced Blackwell chips, which are banned from sale in China by the U.S. DeepSeek reportedly used a method involving installing chips in data centers in other countries, then dismantling and secretly sending them to China to train its new AI models. In response, Nvidia is testing new software that tracks its powerful AI chips, like the Blackwell series, using data center information such as server delays to prevent smuggling into banned markets without relying on GPS. Meanwhile, AI regulation is a key topic across the U.S. New York Governor Kathy Hochul has proposed significant changes to the state's AI safety bill, the RAISE Act. These changes would remove requirements for detailed safety standards and a ban on models with unreasonable risks, while also weakening rules on reporting security incidents and applying to foreign AI models like DeepSeek. An opinion piece argues that Governor Hochul should veto the bill, suggesting that state-level AI rules could hinder American innovation and make the U.S. fall behind China in global AI standards. In Georgia, lawmakers are working on bipartisan bills to enhance child safety on social media and with AI. These proposals include stopping companies from using addictive designs in online platforms and requiring AI developers to test chatbots to ensure they are safe for children. This legislative push comes as Congress has not updated online child protections in nearly 30 years. Separately, GE Vernova Inc. notes that the rising demand for electricity in the U.S. is not solely due to AI and data centers, but also from new factories and electric vehicles. On the application front, global fintech broker ATFX is leveraging KX's AI-driven data platform to improve real-time trading services, providing faster insights and risk management. Similarly, five small and mid-sized law firms are reporting significant benefits from AI tools like CoCounsel Legal, which enables 63% faster document review and doubles the rate of finding relevant cases, giving attorneys up to 12 extra hours weekly for strategic work. For instance, Bochetto & Lentz P.C. used AI to summarize a 50-page report in under an hour. Looking ahead, Microsoft executive Mustafa Suleyman, who leads the company's efforts to build top AI products, has stated that Microsoft will stop developing AI if it ever threatens humanity, emphasizing the crucial role of human oversight for systems like Copilot. OpenAI also plans to introduce an "adult mode" for ChatGPT in 2026, using age estimation technology to direct users to either a "safer" version or the new adult experience, a feature CEO Sam Altman had previously promised.

Key Takeaways

• Chinese AI firm DeepSeek is accused of illegally obtaining Nvidia's advanced Blackwell chips by dismantling them from data centers abroad and shipping them to China.
• Nvidia is developing new software to track its AI chips, including the Blackwell series, to prevent smuggling into banned markets like China.
• New York Governor Kathy Hochul proposed changes to the state's RAISE Act, which would weaken AI safety standards, incident reporting, and application to foreign AI models.
• An opinion piece suggests Governor Hochul should veto the RAISE Act, arguing that state-level AI regulations could harm American innovation and global competitiveness.
• Georgia lawmakers are developing bipartisan bills to make social media and AI chatbots safer for children, including requirements for AI developers to test for child safety.
• ATFX, a global fintech broker, is using KX's AI-driven data platform to enhance real-time trading, risk management, and client reporting.
• U.S. electricity demand is increasing due to AI and data centers, but also significantly from new factories and electric vehicles, according to GE Vernova.
• Five law firms are achieving substantial efficiency gains, such as 63% faster document review and up to 12 extra hours per week for attorneys, by using AI tools like CoCounsel Legal.
• Microsoft executive Mustafa Suleyman stated the company would halt AI development if it threatens humanity, stressing the importance of human oversight for AI products like Copilot.
• OpenAI plans to launch an "adult mode" for ChatGPT in 2026, utilizing age estimation technology to offer different user experiences.

Chinese AI Firm DeepSeek Accused of Smuggling Nvidia Chips

Chinese AI startup DeepSeek faces accusations of illegally obtaining Nvidia's advanced Blackwell chips. The U.S. banned these chips from sale in China to prevent their use in military or surveillance technology. DeepSeek reportedly used a method where chips were installed in data centers in other countries, then dismantled and secretly sent to China. This allowed DeepSeek to train its new AI models despite the restrictions. Nvidia stated it works with authorities to stop illegal diversions. This situation highlights the ongoing tech tensions between the U.S. and China.

Nvidia Tests Software to Track AI Chips and Stop Smuggling

Nvidia Corp. is testing new software to track its powerful AI chips and stop them from being smuggled into banned markets like China. This software uses data from data centers, like server delays, to figure out where chips are located without needing GPS. It is currently being tested on Nvidia's Blackwell series chips, which are important for training AI models. The U.S. banned sales of top-tier Nvidia chips to China in 2022 due to national security concerns. This new tool aims to make it harder for smugglers to illegally move these advanced semiconductors.

New York Governor Hochul Changes State AI Safety Bill

New York Governor Kathy Hochul has suggested major changes to the state's AI safety bill, called the RAISE Act. This bill, created by Senator Andrew Gounardes and Assemblymember Alex Bores, aims to make AI companies more transparent. Hochul's proposed changes remove requirements for detailed safety standards and a ban on models with unreasonable risks. They also weaken rules about reporting security incidents and applying to foreign AI models like DeepSeek. Governor Hochul's office stated that New York wants to lead in responsible AI, protecting consumers while reviewing the legislation.

Opinion Governor Hochul Should Veto New York AI Bill

An opinion piece argues that New York Governor Kathy Hochul should veto the state's AI bill, the RAISE Act. The author believes that having different AI rules in each state will hurt American innovation and make the US fall behind China in setting global AI standards. The article claims the bill has bad rules that help trial lawyers and could make tech companies leave New York for states like Texas or Florida. It also suggests that Congress, not individual states, should create national AI guidelines.

Georgia Lawmakers Plan Bills to Make AI and Social Media Safer for Kids

Georgia lawmakers are working on new bipartisan bills to make social media and AI safer for children. A Senate study committee heard from parents and experts for months before creating its final report. Democratic state Senator Sally Harrell, a co-chair, said the proposals will address concerns about how online platforms affect kids. The planned laws include stopping companies from using addictive designs in social media and games. They also require AI developers to test chatbots to ensure they are safe for children to use. Lawmakers feel they must act because Congress has not updated online child protections in nearly 30 years.

ATFX Boosts Trading with KX AI Data Platform

ATFX, a global fintech broker, is using KX's AI-driven data platform to improve its real-time trading services. This new technology will help ATFX process market data instantly, giving its teams faster insights for trading and risk management. It also allows non-technical users to access complex financial data easily and provides faster, more accurate reports for clients. Jeffrey Siu, ATFX's COO, stated this partnership shows their commitment to using advanced technology for better client service. Ashok Reddy, KX's CEO, added that their platforms will help ATFX use real-time market intelligence and AI to drive innovation.

US Power Demand Rises Beyond Just AI Says GE Vernova

GE Vernova Inc., a major turbine maker, says the growing demand for electricity in the US is not only because of AI and data centers. CEO Scott Strazik explained that new factories and electric vehicles are also big reasons for this increase. The company expects stronger profits in its power and electrification unit. While AI is a significant factor, GE Vernova benefits from being one of only three global makers of large gas turbines. Analysts at JPMorgan Chase & Co. predict the company's shares could reach $1,000.

Five Law Firms See Big Returns Using AI Tools

Five small and mid-sized law firms are seeing great results by using AI tools like CoCounsel Legal. These firms report that AI helps them work smarter, serve clients better, and grow their practices. CoCounsel allows for 63% faster document review and doubles the rate of finding relevant cases. This technology gives attorneys up to 12 extra hours each week for important strategic work and client relationships. For example, Hawaii Disability Legal Services uses AI for faster research, and Bochetto & Lentz P.C. used it to summarize a 50-page report in under an hour. Lawyers emphasize that human judgment remains key, but AI greatly enhances their work.

Microsoft Executive Vows to Stop AI if It Harms Humanity

Microsoft executive Mustafa Suleyman stated that the company will stop developing AI if it ever threatens humanity. He believes this should be a common view in the industry. Suleyman joined Microsoft after his startup, Inflection AI, was acquired, and he now leads Microsoft's efforts to build top AI products. He mentioned that Microsoft can now develop its own AI that could potentially surpass human performance. While current AI systems like Copilot are still improving, Suleyman emphasized that human oversight is crucial.

ChatGPT Will Launch Adult Mode in 2026

OpenAI's ChatGPT is set to introduce an "adult mode" in 2026. CEO Sam Altman had previously promised this feature after some users expressed concerns about the chatbot becoming too restricted. OpenAI plans to use age estimation technology to direct users to the right version of ChatGPT, either a "safer" one for younger users or the new adult experience. This move aims to allow for more "adult" interactions while still providing parental controls.

Sources

• Chinese AI Startup DeepSeek Accused of Smuggling Banned Nvidia Chips
• Nvidia Pilots AI Chip Tracking Software to Curb Smuggling to China
• N.Y. Gov. Kathy Hochul proposes major changes to AI bill
• Gov. Hochul should kill Albany's toxic AI bill
• Georgia lawmakers push bipartisan plan to make social media, AI safer for children
• ATFX Accelerates Real-Time Trading Innovation with KX’s AI-Driven Data Platform for Smarter Trading
• The Boom in US Power Demand Isn’t Just About AI
• ROI Reality check: 5 small and mid sized law firms using AI
• Microsoft Executive Vows to Halt AI Work If It Imperils Humanity
• ChatGPT's 'Adult Mode' Is Coming in 2026