California has become the first state to ban employers from relying solely on AI to fire or discipline workers. Governor Gavin Newsom signed the No Robo Bosses Act, authored by Democratic State Senator Jerry McNerney, along with two other workplace AI bills on September 30, 2026. Under the new law, companies that primarily use AI for termination decisions must have a human corroborate those decisions using manager evaluations, peer reviews, and personnel files. Employees must also receive written notice that AI played a primary role and be given a human contact for questions.
The legislation targets several areas of AI use in the workplace. Assembly Bill 1883 bans AI tools that predict employees' emotional states or collect neural data, while Assembly Bill 1331 prohibits surveillance in bathrooms. Senate Bill 947 specifically addresses the sole use of automated systems for discipline or firing. Violations can carry penalties of up to $500 per violation. Newsom also signed an executive order requiring state agencies to use the term "artificial intelligence" instead of "AI." He criticized the Trump administration for failing to pass comprehensive federal AI regulations.
Not all labor-focused AI bills were signed into law. The measure curbing AI in firing workers was weakened during the legislative process and no longer covers contractors. California Labor Federation president Lorena Gonzalez said more work remains, especially regarding AI regulation in healthcare.
In other AI news, Google has launched Gemini 4 Argon, its flagship model, starting with trusted cybersecurity partners. Google said the model scored well on benchmarks, including beating OpenAI's Astra on a security test. However, some insiders report that Gemini 4 struggles with coding and front-end design tasks. Versions of Gemini now power products used by over 1 billion users, though some employees believe models from Anthropic and OpenAI are improving faster.
Nvidia is facing scrutiny on Wall Street over its large chip financing plan. Lenders are reportedly seeking higher guarantees than Nvidia originally proposed, despite the company's leading position in AI chip technology. Banks and investors are debating how long revenue from Nvidia chips and related infrastructure will remain strong.
AI-focused real estate startups are drawing significant capital. Kahua, an AI construction platform, raised $250 million from Bain Capital, while EliseAI, a multifamily AI startup, raised $350 million. Kahua has supported over $400 billion in capital programs for clients including the Port Authority of New York and New Jersey and JFK Airport. EliseAI's automation platform serves more than 500 owners and operators in the multifamily sector.
GTT Communications introduced Defense Halo, an AI-powered network defense platform that creates real-time models of customer networks to detect vulnerabilities and hunt for threats. Analyst Amy DeCarlo noted it removes blind spots found in traditional security products. Separately, a new Baseline Security Checklist helps CPAs manage cybersecurity risks from generative and agentic AI tools across 12 key areas.
Key Takeaways
- California signed the No Robo Bosses Act, banning employers from solely using AI to fire or discipline workers, with human review now required for such decisions.
- Three AI workplace bills were signed by Governor Newsom on September 30, 2026, including bans on AI emotional prediction and bathroom surveillance, with violations penalized up to $500 each.
- The signed AI firing law was weakened during the legislative process and no longer covers contractors, according to labor leaders.
- Google launched Gemini 4 Argon, claiming it beat OpenAI's Astra on a security benchmark, though some insiders say it struggles with coding and front-end design.
- Nvidia faces Wall Street debate over its chip financing plan, with lenders seeking higher guarantees than originally proposed.
- Kahua raised $250 million from Bain Capital and EliseAI raised $350 million, reflecting heavy investment in AI real estate technology.
- GTT Communications launched Defense Halo, an AI-powered platform that models customer networks in real time to detect security threats and vulnerabilities.
- A new Baseline Security Checklist helps CPAs evaluate cybersecurity maturity across 12 areas to address risks from generative and agentic AI tools.
- Newsom signed an executive order requiring state agencies to use "artificial intelligence" instead of "AI" and criticized the lack of federal AI regulation.
- Japanese audit firms are merging to meet new partner requirements and invest in AI as a five-year grace period on client fee limits ends.
California bans AI robo bosses in new workplace law
California Governor Gavin Newsom signed the No Robo Bosses Act, making California the first state to ban employers from solely using AI to fire or discipline workers. The law requires human review when AI is the primary tool for termination or discipline decisions. Employers must also notify employees when AI was primarily used and provide a human contact for questions. The bill was authored by Democratic State Senator Jerry McNerney.
California governor signs three AI workplace surveillance bills
On September 30, 2026, Governor Gavin Newsom signed three bills targeting AI in the workplace. Assembly Bill 1883 bans AI tools that predict employees' emotional states or collect neural data. Assembly Bill 1331 prohibits workplace surveillance in bathrooms. Senate Bill 947, the No Robo Bosses Act, prevents employers from relying solely on automated systems for discipline or firing. Violations can carry penalties of up to $500 per violation.
Newsom signs law barring sole use of AI in hiring and firing
California Governor Gavin Newsom signed a law barring employers from solely using artificial intelligence to hire, discipline, or fire workers. The bill requires businesses that primarily rely on AI for termination decisions to have a human corroborate those decisions using manager evaluations, peer reviews, and personnel files. Employees must receive written notice that AI was primarily used and a human contact to discuss the decision. The bill was authored by Democratic State Senator Jerry McNerney.
Newsom signs some but not all AI labor bills
Governor Gavin Newsom signed several bills protecting workers from AI but did not sign all the measures labor leaders pushed for. The signed laws limit how workers can be terminated or surveilled by artificial intelligence. Labor leaders, including California Labor Federation president Lorena Gonzalez, said more work is needed, especially to regulate AI in healthcare. The law curbing AI in firing workers was weakened during the legislative process and no longer covers contractors.
California signs laws to protect workers from AI risks
Governor Gavin Newsom signed laws protecting workers from AI risks, including bans on using AI to predict a worker's emotional state through biometric data. The laws also require written notices to workers when AI is responsible for mass layoffs and ban employers from relying on AI to decide to fire someone. Newsom also signed an executive order requiring state agencies to use the term artificial intelligence instead of AI. He criticized the Trump administration for not passing comprehensive federal AI regulations.
California bans AI-only firings with new No Robo Bosses Act
California Gov. Gavin Newsom signed SB 947, the No Robo Bosses Act of 2026. The law stops employers from using AI alone to fire or discipline workers. A human must now check any AI-based decision. Workers must be told AI played a main role and given a contact person. The bill was authored by state Sen. Jerry McNerney and sponsored by the California Federation of Labor Unions, AFL-CIO.
Alphabet shares dip over doubts about Gemini 4 AI model
Alphabet stock fell after a report showed internal doubts about Gemini 4. Employees say Gemini 4 does well on tests but struggles with real tasks like coding. Google disputed this and pointed to Google DeepMind head Koray Kavukcuoglu, who said he was encouraged by the model. Some employees think models from Anthropic and OpenAI are improving faster. Other employees believe Gemini 4 has caught up.
Google rolls out Gemini 4 Argon amid internal skepticism
Google launched Gemini 4 Argon, its flagship AI model, starting with trusted cybersecurity partners. The company said Gemini 4 scored well on benchmarks, including beating OpenAI's Astra on a security test. However, some insiders say it struggles with coding and front-end design. Google said it is inaccurate to call the model an underperformer. Versions of Gemini now power products used by over 1 billion users.
Japan audit firms merge to meet partner rules and invest in AI
Smaller Japanese audit firms are using mergers to meet new partner requirements and invest in AI. Two mergers closed on Oct. 1. Firms need more credentialed accountants as partners to vet books of public companies. A five-year grace period is also ending for a rule that limited auditing fees from one client to 15% of total revenue.
Japan audit firms merge to meet partner rules and invest in AI
Smaller Japanese audit firms are using mergers to meet new partner requirements and invest in AI. Two mergers closed on Oct. 1. Firms need more credentialed accountants as partners to vet books of public companies. A five-year grace period is also ending for a rule that limited auditing fees from one client to 15% of total revenue.
New docuseries uses AI to tell Aileen Wuornos' story
A new docuseries called Unmasking a Monster: Aileen Wuornos uses AI technology to bring the story of serial killer Aileen Wuornos to life. Wuornos confessed to killing seven men between 1989 and 1990 and was executed in 2002. Executive producer Stephen Dost and actress Morag Peacock discussed the series on Forbes True Crime.
Wall Street questions Nvidia chip financing plan
Nvidia's large financing plan has caused debate on Wall Street about the value of its advanced chips and related infrastructure. Some lenders want higher guarantees than Nvidia originally proposed, even with the company's leading AI chip technology. Banks and investors are trying to figure out how long revenue from Nvidia chips will last.
GTT launches AI-powered Defense Halo security platform
GTT Communications introduced Defense Halo, a new network defense platform that uses AI to detect and fix security threats. The platform creates a real-time model of a customer's network to find vulnerabilities and hunt for threats. Fletcher Keister, GTT's Chief Product and Technology Officer, said the tool addresses risks from undetected vulnerabilities. Analyst Amy DeCarlo said it removes blind spots found in traditional security products.
New checklist helps CPAs manage AI cyber risks
A new Baseline Security Checklist was created to help CPAs and accounting firms manage cybersecurity risks from AI. The checklist was developed by Sanjay Chadha of SAV Associates to evaluate a firm's cybersecurity maturity across 12 key areas. Each area is scored with up to 75 points based on five to 10 checkpoints. The tool addresses growing threats from generative and agentic AI tools used by attackers.
AI real estate startups raise hundreds of millions
AI-focused real estate technology companies are attracting major investments. Kahua, an AI construction platform, raised $250 million from Bain Capital, while EliseAI, a multifamily AI startup, raised $350 million. Kahua has supported over $400 billion in capital programs for clients including the Port Authority of New York and New Jersey and JFK Airport. EliseAI's automation platform serves more than 500 owners and operators in the multifamily sector.
3 US AI Stocks Tied to Miami's Tech and Finance Growth
Miami is becoming a major tech and finance hub, boosted by Ken Griffin's US$3b gift to Carnegie Mellon. Globant is highlighted as one US Technology and AI stock linked to this shift. The company builds AI-powered digital platforms for large enterprises and earns about US$2.5b in revenue. The article also notes a full screener that found 41 more US-listed tech companies with similar stories.
Study Maps Self-Distillation Methods for Language Models
Researchers studied self-distillation with privileged context, where a language model teaches a copy of itself token by token. They tested all combinations of three key axes using Qwen2.5-7B and Ministral-3-3B models across 1,200 adaptation runs. The findings show that rollout source, teacher coupling, and KL direction each affect learning and retention differently. A controlled model was proposed to explain these trade-offs.
Sources
- California Gov. Gavin Newsom bans AI 'robo bosses' in landmark state law, reversing his earlier veto
- California Governor Signs 3 Bills Targeting AI and Workplace Surveillance
- Newsom signs law barring sole use of AI in hiring, firing decisions
- On AI, Newsom gives labor only some of what it demanded
- California Gov. Gavin Newsom signs laws to protect workers from AI risks
- California bans employers from using AI alone to fire workers
- Alphabet stock slips on report of internal doubts over Gemini 4
- Google’s latest launch puts its AI ambitions to the test
- Japan’s Audit Firms Merge to Meet Partner Floors, Invest in AI
- Japan’s Audit Firms Merge to Meet Partner Floors, Invest in AI
- Inside The New Docuseries That Uses AI To Bring Serial Killer Aileen Wuornos’ Story To Life
- Nvidia's bet that its chips can finance the AI boom gets a Wall Street reality check
- GTT targets network blind spots with AI-native security
- New checklist helps CPAs manage AI cyber risks
- AI real estate firms reel in millions of dollars
- 3 US AI Stocks Linked To Miami’s Growing Tech And Finance Hub
- Disentangling Self-Distillation: Measuring and Modeling Acquisition and Retention
Comments
Please log in to post a comment.