deepseek, nvidia and google Updates

Recent developments in the AI sector highlight both its rapid expansion and the growing challenges it presents, from regulatory hurdles to ethical concerns. Chinese AI startup DeepSeek faces scrutiny over reports alleging its use of banned Nvidia Blackwell chips, which were supposedly smuggled into China after being installed in data centers in other countries. Nvidia, however, denies finding any proof of such smuggling, calling the idea "farfetched." The US maintains a ban on exporting these advanced chips to China, aiming to preserve its lead in AI technology. DeepSeek previously gained attention for its R1 reasoning model and also collaborates with Chinese chipmakers on hardware development. Meanwhile, Google has rolled out significant security updates for Chrome's AI browsing features to combat "indirect prompt injection attacks." This new system incorporates a "User Alignment Critic" to monitor AI actions without direct web content access and "Agent Origin Sets" to restrict AI access to specific website parts. Chrome's AI will also seek user permission before executing sensitive tasks like accessing banking sites or making purchases. Google is offering a $20,000 reward to researchers who can successfully breach these new security measures, aiming to rigorously test their effectiveness. Globally, AI adoption continues to reshape various industries. India's Bureau of Indian Standards (BIS) announced 12 new standards on December 10, 2025, covering AI data quality, accountability, and bias, alongside secure financial cryptographic devices. In the financial sector, a Bloomberg Intelligence survey indicates that AI adoption on Wall Street will likely lead to an initial increase in hiring, with two-thirds of companies expecting staff growth and over 70% anticipating higher operating costs in the next three years, though productivity is expected to rise even faster. Bitget, a cryptocurrency exchange, launched GetAgent, an AI-powered trading tool designed to assist investors with market analysis and strategy. Bitget CEO Gracy Chen describes GetAgent as a "game plan" for traders, converting data into actionable strategies. AI's integration also extends to agriculture, where tractor maker CNH is incorporating AI into its equipment and operations. CNH, with Microsoft as a key technology partner, uses AI to help software engineers reduce documentation time by 60% for precision agriculture and enables spraying systems to cut herbicide use by 80% by accurately identifying weeds. They also deployed an "AI Tech Assistant" for dealers in early 2025. However, the rapid spread of AI also brings new challenges, as evidenced by a class action lawsuit filed on November 26 in San Diego against healthcare AI recording practices, highlighting risks for businesses using AI to record customer conversations. Academic integrity also faces new threats, with a professor noting students misusing generative AI to create summaries of non-existent articles and authors. In the realm of AI research and competition, Associate Professor Dongjin Song from UConn was named a Senior Member of the Association for the Advancement of Artificial Intelligence (AAAI) on December 10, 2025, for his work in machine learning and agentic AI. Furthermore, Aster DEX launched a $200,000 Human vs AI trading competition from December 9 to December 23, pitting 70 human traders against 30 AI agents with fully funded accounts, aiming to test the performance of human intuition against advanced AI systems in live market conditions.

Key Takeaways

• Chinese AI startup DeepSeek is reportedly using banned Nvidia Blackwell chips, allegedly smuggled into China, though Nvidia denies finding any proof.
• Google introduced new AI security features for Chrome, including a "User Alignment Critic" and "Agent Origin Sets," and offers a $20,000 reward for breaching them.
• India's Bureau of Indian Standards (BIS) announced 12 new standards on December 10, 2025, covering AI data quality, accountability, and secure financial cryptographic devices.
• Wall Street expects AI adoption to initially boost hiring and operating costs, with two-thirds of companies anticipating staff increases.
• Bitget launched GetAgent, an AI-powered crypto trading tool, with CEO Gracy Chen emphasizing its role in converting data into trading strategies.
• CNH, in partnership with Microsoft, integrates AI into farming equipment, reducing software documentation time by 60% and herbicide use by 80% with AI-enabled spraying.
• A class action lawsuit was filed on November 26 in San Diego, targeting healthcare AI recording practices, highlighting legal risks for businesses.
• Academic integrity faces challenges as students misuse generative AI to create summaries of non-existent research articles.
• Associate Professor Dongjin Song from UConn was named an AAAI Senior Member on December 10, 2025, for his contributions to machine learning and agentic AI.
• Aster DEX is hosting a $200,000 Human vs AI trading competition from December 9 to December 23, featuring 70 human traders against 30 AI agents.

China's DeepSeek uses banned Nvidia chips for AI

A report by The Information says Chinese AI startup DeepSeek used banned Nvidia Blackwell chips for a new AI model. These chips were reportedly smuggled into China through other countries. Nvidia stated it has not found any proof of such smuggling and called the idea "farfetched." The US bans selling these advanced chips to China. DeepSeek previously released a large language model and also works with Chinese chipmakers to develop its own hardware.

China's DeepSeek uses banned Nvidia chips for AI

A report states that Chinese AI company DeepSeek used banned Nvidia Blackwell chips for its AI model. The Information reported that these chips were smuggled into China. They were first installed in data centers in other countries, then taken apart and sent to China after inspections.

Nvidia denies DeepSeek uses banned Blackwell AI chips

Nvidia has denied a report claiming that Chinese AI startup DeepSeek uses its banned Blackwell AI chips. An Nvidia spokesperson stated they have not found any proof or received tips about such smuggling operations. The US banned the export of Nvidia's advanced Blackwell chips to China to maintain its lead in AI technology. DeepSeek gained attention earlier this year with its R1 reasoning model, which performed well and was developed at a low cost.

Google Chrome gets new AI security to block hackers

Google introduced a major security update for Chrome's AI browsing features to stop hackers. This new system aims to block "indirect prompt injection attacks" where malicious instructions are hidden in web content. Key features include a "User Alignment Critic" that checks AI actions without seeing web content, and "Agent Origin Sets" that limit AI access to certain website parts. Chrome's AI will also ask for user permission before sensitive actions like accessing banking sites or making purchases. Google offers a $20,000 reward to researchers who can breach these new security measures.

Google offers $20,000 reward for Chrome AI security breach

Google is offering a $20,000 reward to anyone who can breach the new security features for Chrome's AI agents. This bounty aims to test the defenses against "indirect prompt injection" attacks, where hidden instructions could trick AI into unauthorized actions. The new system includes a "User Alignment Critic" that acts as an AI watchdog, checking every action the main AI plans to take. Google also added "Agent Origin Sets" to create strict boundaries for AI access to websites and requires human permission for sensitive tasks like financial transactions or accessing private data.

India rolls out 12 new standards for AI and security

The Bureau of Indian Standards BIS announced 12 new standards on December 10, 2025. These standards cover several important areas, including artificial intelligence, secure financial cryptographic devices, and audit data systems. New AI data-quality standards are especially important for improving accountability and reducing bias in regulated industries. BIS also introduced a new standard for secure cryptographic devices used in retail financial services, matching international guidelines.

Professor shares lessons from students misusing AI

A professor shared experiences with students improperly using generative AI in a general psychology course. Students were assigned to summarize and reflect on current research articles from specific journals. One student used AI to generate article summaries, which led to the use of non-existent articles by fake authors like "Smith" and "Doe." The professor discovered this when checking references, realizing students might write their own papers but rely on AI-created summaries of fake sources. This highlights challenges for instructors in maintaining academic integrity with new AI tools.

Wall Street AI adoption may boost hiring initially

A Bloomberg Intelligence survey suggests that AI adoption in financial services will likely lead to more hiring at first, not immediate job cuts. Two-thirds of companies expect staff numbers to increase, and over 70% anticipate higher operating costs in the next three years. However, most also believe productivity will rise more quickly. Experts say the early phase of AI is about building new capabilities rather than reducing costs. In the long run, AI is expected to significantly impact the industry, with banks potentially benefiting greatly by 2026.

Bitget GetAgent uses AI to boost crypto trading

Bitget launched GetAgent, a new AI-powered trading tool designed to help cryptocurrency investors. GetAgent uses an adaptable system to answer user questions and provides easy access to trade previews, research, and market analysis. Recent updates include a smoother interface and more daily queries for all users, with the free Basic tier now getting 20 queries per day. This tool aims to help traders improve their strategies and make better decisions in the fast-changing crypto market. Bitget CEO Gracy Chen compares GetAgent to a game plan for traders, turning data into strategies.

UConn Professor Dongjin Song named AAAI Senior Member

On December 10, 2025, Associate Professor Dongjin Song from UConn received Senior Member status from the Association for the Advancement of Artificial Intelligence AAAI. He will be recognized at the 40th Annual AAAI Conference in Singapore in January 2026. Professor Song specializes in machine learning, data science, and deep learning, with a focus on areas like time series analysis and agentic AI. His influential research includes papers such as DARNN and HetGNN. AAAI, founded in 1979, is a nonprofit group dedicated to advancing safe and responsible AI for society.

New lawsuit targets healthcare AI recording practices

A new class action lawsuit was filed on November 26 in San Diego, targeting healthcare AI recordings. This lawsuit highlights a major risk for any business that uses AI tools to listen, record, or summarize conversations with customers or patients. The case involves Sharp, a healthcare provider. Businesses should consider steps to limit their exposure to similar legal challenges.

CNH brings more AI to farming operations

Tractor maker CNH is actively integrating artificial intelligence into its farming equipment and business operations. Jay Schroeder, CNH's Chief Technology Officer, notes the company is focused on measuring the return on investment for AI. AI is already helping software engineers reduce documentation time by 60% for precision agriculture technology. Additionally, AI-enabled spraying systems can cut herbicide use by 80% by accurately identifying weeds. CNH also uses AI for tasks like drafting contracts and quickly generating field reports. The company's main technology partner is Microsoft, and they launched an "AI Tech Assistant" for dealers in early 2025.

Humans and AI battle in $200,000 trading contest

Aster DEX launched a Human vs AI trading competition with a $200,000 prize pool, running from December 9 to December 23. Seventy human traders are competing against 30 AI trading agents. Both humans and AI receive fully funded accounts, meaning participants trade with no personal risk. This event aims to test if human intuition can outperform advanced AI systems in live market conditions. Aster designed the contest as a major experiment into the future of trading intelligence.

Sources

• China’s DeepSeek Uses Banned Nvidia Chips for AI Model, The Information Says
• China’s DeepSeek Uses Banned Nvidia Chips for AI Model, Report Says
• Nvidia refutes report that China's DeepSeek is using its banned Blackwell AI chips
• Google Chrome's New AI Security Aims to Stop Hackers Cold
• Google Offers $20K Bounty for Chrome AI Security Breach
• Govt rolls out 12 new standards across AI, payments security, audit data and industrial systems
• Lessons Learned from Students Using AI Inappropriately in My Class
• Wall Street’s AI Adoption Is Set to Drive Hiring Boom, For Now
• Bitget's GetAgent Transforms AI Trading for the Discerning Investor
• UConn Professor Receives Senior Member Status with AAAI
• New Class Action Targets Healthcare AI Recordings: 6 Steps All Businesses Should Consider to Limit Exposure
• Inside tractor maker CNH’s push to bring more artificial intelligence to the farm
• Humans vs AI Go Head-to-Head in Aster’s $200,000 Trading Showdown