New York Proposes $1 Million Daily Fines for AI Violations

Key Takeaways

• Proposed New York RAISE Act 2.0 targets frontier-model developers spending more than $1 billion over three years and would impose fines of up to $1 million per day for a first violation and $3 million per day for later violations; a 30-day repeat violation could mean up to $90 million.
• New York’s AI safety center at SUNY Binghamton launched Nov. 11, 2026, with more than 80 scientists, a $30 million philanthropic gift and an initial $25 million SUNY investment.
• Microsoft says its Quincy data center employs about 700 people and that it pays for electricity and replenishes more water than it uses; Grant County electricity rates rose about 55% over 20 years as data-center development expanded.
• Illumio found 91% of surveyed Singapore organisations faced an AI-related security issue or warning sign, and only 43% could audit every AI agent action.
• UC Berkeley and NVIDIA released RoboSimStudio, an open-source tool using AI coding agents to create and refine robot simulation programs.
• Lawmakers’ calls for Nvidia CEO Jensen Huang to resign could sharpen congressional debate over advanced-chip export controls, including exports to China.
• BARS is designed to assess whether repeated LLM answers are reliable within a set budget, with options to stop early, return a majority answer or refuse when it cannot confirm reliability.
• IIT Roorkee and IIT Hyderabad announced AI-related courses, including Roorkee’s Rs 8,000 six-week programmes and Hyderabad’s three-month Forward Deployed Engineering course beginning Dec. 5, 2026.
• Bottomline says customers in payments want AI guidance grounded in industry expertise, safeguards and trust, especially around sensitive data, fraud and compliance.
• A Tuya stock analysis sets a $3.26 target ahead of Q3 2026 earnings, citing 22% growth in consumer recurring-service revenue while flagging uncertain AI monetisation and other risks.

New York bill would raise AI safety fines to $3 million a day

New York’s proposed RAISE Act 2.0 would fine frontier AI developers up to $1 million per day for a first safety violation and up to $3 million per day for later violations. A repeat violation lasting 30 days could lead to a fine of up to $90 million, though the state attorney general would need to sue to collect it. The bill would cover developers that spent more than $1 billion on frontier models over three years and require yearly outside safety audits, reports every three months on AI-run research, and a senior officer to certify filings. The original law takes effect Jan. 1, 2027. The article also notes that 89% of enterprises use agentic AI, while 78% would limit these tools to reading company directories, according to PYMNTS Intelligence.

New York opens statewide AI safety research center at SUNY Binghamton

New York launched the New York Centre for AI Responsibility and Research at SUNY Binghamton on Nov. 11, 2026. The center will test advanced AI models, study their risks, develop safety tools and train specialists. Interim co-directors Greg Morrisett of Cornell Tech and Jeannette Wing of Columbia University will guide more than 80 scientists from institutions across New York State. Funding includes a $30 million philanthropic gift led by Binghamton alumnus Tom Secunda and an initial $25 million research investment from SUNY. The center will also study responsible AI use in medicine, scientific discovery and critical infrastructure.

New York opens AI safety center before RAISE Act takes effect

New York Gov. Kathy Hochul announced the New York Center for AI Responsibility and Research at SUNY Binghamton, bringing together more than 80 scientists to test AI systems and train safety experts. The center will develop tools and standards to assess advanced systems, study their risks and support responsible AI use in medicine, research and critical infrastructure. Its work will support the RAISE Act, which takes effect Jan. 1. Frontier AI developers operating in New York must register by Jan. 1, publish safety frameworks and model transparency reports, and report critical safety incidents within 72 hours. The center can advise state officials and review public disclosures, but researchers will not have access to companies’ private models under development.

Microsoft data center brings growth and challenges to Quincy, Washington

Microsoft has operated its Columbia Data Center in Quincy, Washington, for 20 years. Mayor Paul Worley said tax revenue from data centers helped fund a new high school and hospital and improve other services. Microsoft President Brad Smith said the company aims to put communities first by paying for its electricity and replenishing more water than it uses. Quincy’s access to Columbia River hydropower has drawn seven other companies, and roughly two dozen data centers have entered the area. Grant County’s electricity rates rose about 55% over 20 years, while Microsoft says it employs about 700 people in a town of roughly 8,000 residents.

AI data center growth raises investor concerns over energy and the environment

XPRIZE founder Peter Diamandis discussed concerns about AI data centers’ environmental effects, rising energy needs and power limits in an appearance on “Making Money.” He also spoke about developments in artificial intelligence and semiconductor technology. The discussion included how programmable biology could change healthcare and disease treatment.

BARS sets budget-aware rules for reliable majority answers

Researchers propose BARS, a method for deciding when repeated answers from a large language model are reliable enough to use. It checks the gap between the two most common answers and can stop early, return a majority vote within a set budget, or refuse when it cannot confirm the result. A refusal means the rule cannot establish a reliable answer within its budget, not that the question has no answer. The method offers a certified mode with proven error bounds and a practical mode based on empirical calibration. Tests on synthetic data across thousands of seeds showed the method responded to how difficult each case was.

IIT Roorkee and IIT Hyderabad announce new AI training programmes

IIT Roorkee and IIT Hyderabad have announced AI programmes for different learners. IIT Roorkee will offer five online, six-week training and internship courses through its EICT Academy, with topics including Generative AI, Machine Learning, Data Science, web development, and Internet of Things and Edge AI. The fee is Rs 8,000, applications are due by November 30, and batches begin on December 14 and January 4. IIT Hyderabad will offer a three-month Forward Deployed Engineering programme for software and AI professionals, starting December 5, 2026, with online classes and optional campus immersion. Both programmes provide a certificate from the institute.

Bottomline says AI needs expertise and trust to deliver customer value

Bottomline says customers want help from people who understand their businesses, not just the newest AI tools. In payments and financial operations, customers are interested in AI but want guidance on sensitive data, workflows, fraud, compliance, and automation risks. The company says businesses also want strong safeguards and clear explanations of how AI can improve finance operations. Bottomline argues that AI makes technology easier to create but does not replace knowledge, experience, or trust. It plans to pair AI investment with industry expertise, responsible use, and trusted customer relationships.

Tuya stock outlook focuses on cash value and AI potential

An analysis rates Tuya bullish and sets a target of $3.26 per share ahead of the company’s Q3 2026 earnings. It says the stock trades at a deep discount and at net cash value. Key areas to watch include execution in PaaS, Smart Home and Robotics, and AI applications, with PaaS and Smart Home showing strong year over year revenue growth. Slow growth in AI applications is linked to Tuya’s strategic exit from B2B, while consumer-facing recurring service revenue grew 22% last quarter. Risks include commoditized device connectivity, exposure to China’s economy, and the possibility that AI products do not generate significant revenue despite continued investment.

Singapore leads surveyed markets in AI-related security incidents

Research from Illumio found that 91% of surveyed Singapore organisations faced an AI-related security issue or warning sign in the past 12 months, compared with 76% globally. The study surveyed 1,100 senior IT security decision-makers and found that 74% of Singapore organisations now let AI make decisions or take actions that needed human approval a year earlier, while only 43% can audit every AI agent action. Four in 10 organisations said an AI tool connected to systems without approval, and 35% reported an unauthorised action triggered by AI. Illumio’s Andrew Kay said organisations should limit what AI can access and enforce those limits through Zero Trust, rather than trying to predict every action it might take.

Berkeley and NVIDIA develop AI agents to program robot simulations

UC Berkeley and NVIDIA released RoboSimStudio, an open-source tool for creating robotic simulation environments. AI coding agents can write, test, and refine readable robot programs. The approach aims to reduce the weeks or months that programming and troubleshooting industrial robots can take. Ken Goldberg of UC Berkeley said the agents can help engineers review, fine-tune, and verify the programs without collecting massive amounts of data or hand-tuning every behavior.

Lawmakers call for Jensen Huang to resign as Nvidia export controls loom

Some lawmakers are calling for Nvidia CEO Jensen Huang to resign amid a scandal. The situation could give Congress a reason to advance bills in this year’s National Defense Authorization Act that would tighten export controls on advanced AI chips. Nvidia’s next steps and whether Huang remains CEO could affect the company’s relationship with Washington. They could also influence whether lawmakers crack down on chip exports to China.

Sources

• New York Proposes $3 Million Daily Fines for AI Safety Violations
• New York launches frontier AI centre at SUNY Binghamton
• New York Launches AI Safety Center Ahead of RAISE Act Deadline
• Microsoft data center in Washington state brings "a kind of boom" despite growing AI pushback
• Blue wave raises investor concerns over AI data center expansion
• BARS: Budget-Aware Sequential Sampling with Refusal for Majority Identification under Unknown Distributions
• IIT Roorkee, IIT Hyderabad announce new AI courses: Check short-term training programmes, fees, eligibility and how to apply
• Bottomline Sees Expertise Turning AI Potential Into Customer Value
• Tuya: Trading At Net Cash With A Free Call Option On Agentic AI (Q3 Preview)
• Singapore tops global rankings for AI-related security incidents
• Can AI agents program industrial robots without training data? Berkeley says yes
• Jensen Huang affair in DC: Resignation calls and moves to tighten Nvidia China exports