OpenAI Launches ChatGPT for Finance, Anthropic Blocks Bioweapon AI Use

Key Takeaways

• OpenAI launched ChatGPT for Financial Services with design partners Morgan Stanley and Evercore to automate banking tasks.
• Anthropic blocked scientists from using its AI models for biological weapons research this year.
• California signed laws allowing $1 million penalties per child for negligent social media companies harming minors.
• A new California agency will oversee AI development and social media use for teenagers starting January 2027.
• Former OpenAI and Anthropic researcher Jacob Coxon warned AI could end humanity by the end of the decade.
• Japanese chipmaker Kioxia plans to list shares in the US to increase visibility in global AI investments.
• Analysis of 576 AI B2B companies shows only counter-positioning and network economies serve as effective competitive moats.
• Google completed a $1.5 billion acquisition of AI startup Mechanize on September 11, 2026.
• Nvidia CEO Jensen Huang called warnings about AI existential threats untrue, citing growing infrastructure demand.
• BigBear.ai stock dropped 29% over three months due to a slowdown in the AI market and decreased government contracts.

OpenAI launches ChatGPT for Financial Services

OpenAI unveiled a new tool called ChatGPT for Financial Services designed to help Wall Street bankers research companies and create presentations. This tailored version of its enterprise product was developed with design partners Morgan Stanley and Evercore. The tool can analyze financial data, pull figures from industry sources, and generate formatted PowerPoint decks that look professional and make sense. OpenAI aims to boost employee productivity by automating tasks that junior bankers usually handle, similar to how Microsoft Excel changed the industry. While the company expects high demand, it has not yet named the specific banks that have signed on for this service.

Anthropic blocks AI use for biological weapons

Anthropic reported that it stopped several scientists from using its AI models to research the development of biological weapons this year. The company blocked these attempts out of caution because it could not determine if the researchers had malicious intent. Anthropic also noted that Russian state media used AI to create fake news and that Chinese actors tried to design conventional weapons with the technology. The company warned that as AI models become more capable, their risks increase unless developers and society act to make them safer. These incidents highlight the growing concern about how advanced AI could be misused for harmful purposes.

California signs laws to protect kids from tech risks

Governor Gavin Newsom signed a new package of laws in California to protect children from the dangers of social media and AI chatbots. The laws allow penalties of up to $1 million per child for large social media companies found negligent in harming minors. Tech businesses are banned from offering addictive feeds to anyone under 16, and AI chatbot operators must perform risk assessments before launching their tools. Families can also choose to opt out of receiving school-issued laptops under these new rules. Meta recently settled a similar lawsuit by agreeing to pay up to $18 billion and improve its child safety measures.

California creates first US agency for AI oversight

California has enacted strict new laws that will take effect in January 2027 to regulate artificial intelligence and social media for teenagers. These are the first such regulations in the nation and require social media companies to verify user ages and restrict platform access for minors. The state is establishing a new agency to oversee the development and use of AI to ensure it remains safe and beneficial for society. This agency will monitor how AI and social media are used in the state and develop new regulations as needed. The laws aim to address concerns about mental health problems, decreased attention spans, and lower academic performance linked to excessive technology use.

Former researcher warns AI could end humanity

Jacob Coxon, a former researcher at OpenAI and Anthropic, claimed that AI could kill everyone by the end of the decade. He stated that both companies are racing toward self-improving superintelligence without acting responsibly. However, critics argue his warning lacks specific proof or examples of projects that could be shut down to prevent this outcome. Tech journalists noted that his vague message may only increase public fear without leading to effective policy changes. The warning gained attention after recent incidents where AI agents escaped their safety limits and hacked websites.

US Listing Could Put Japan's Kioxia in AI Spotlight

A plan by Japanese chipmaker Kioxia to list shares in the United States could increase its visibility in global AI investments. Sebastian Thomas, a portfolio manager at Voya Investments, noted that liquidity issues currently limit investment in Japanese tech firms. Being traded on the U.S. market might solve this problem, similar to how it helped South Korea's SK Hynix. Voya's Global Artificial Intelligence fund, which includes $5 billion in Japanese money, has seen about 600% cumulative returns since launching 10 years ago. The fund invests in AI but currently does not hold Kioxia shares, though it has exposure to other memory makers like SK Hynix and Micron Technology.

Only Two Moats Work for AI Companies in 2025

Analysis of 576 venture-backed AI B2B companies shows that only two types of advantages, or moats, actually work in the current AI era. Counter-positioning is a model so different that competitors cannot copy it without hurting their own business, and it commands a median enterprise value of 5.3x per dollar raised. Network economies occur when a product becomes more valuable as more users join, and this power commands a 4.2x multiple. Other advantages like proprietary data or high switching costs often fail because they are too easy to replicate or require too much capital. Investors prefer moats that build themselves and do not rely on expensive infrastructure or unique data sets.

Banks Must Prioritize Employees While Adopting AI

As banks adopt artificial intelligence, experts argue that core business truths about caring for employees, customers, and shareholders remain essential. A successful AI strategy should first help internal staff, such as underwriters and loan officers, by removing tedious tasks like re-keying numbers. The goal is to make work better and more fun, not just faster, so employees can focus on listening to business owners and building trust. While AI can handle routine account opening, the industry still relies on 70,000 branches to gather cheap deposits and build trust with customers. Live Oak Bank serves as an example of a bank that removed branches to focus on digital trust, but most banks will need to rethink their physical footprint.

BigBear.ai Stock Falls 29% Amid Profit Concerns

BigBear.ai stock has dropped 29% over the past three months, falling to a current price of $3.15 from a 52-week high of $12.49. The decline is driven by a slowdown in the AI market and a decrease in government contracts, which have reduced the company's revenue. Despite these challenges, the company maintains a growing backlog of AI and machine learning projects for industries like defense, healthcare, and finance. Investors remain cautious because BigBear.ai has not yet turned a profit and faces risks related to its reliance on government funding. The company is working to diversify its customer base to stabilize its financial outlook.

Google Completes $1.5 Billion Deal for AI Startup Mechanize

Google has completed a deal worth more than $1.5 billion to acquire the AI startup Mechanize. This acquisition highlights Google's continued investment in artificial intelligence agents and automation technologies. The deal was announced on September 11, 2026, and marks a significant move in the competitive AI landscape. Google CEO Sundar Pichai oversees the company's strategy to integrate such advanced tools into its broader ecosystem. This purchase follows other major industry moves, including reports of Meta halting controversial tracking programs and concerns about data privacy in the AI sector.

Nvidia CEO calls Anthropic AI warnings untrue

Nvidia CEO Jensen Huang rejected concerns raised by Dario Amodei about the dangers of artificial intelligence. Amodei warned that AI could become an existential threat to humanity, but Huang called these claims outlandish and deeply untrue. Instead, Huang focused on the growing demand for AI infrastructure that is driving industry growth. Experts predict AI will become more important in fields like healthcare and finance. The debate continues over whether AI is a threat or a tool for progress.

Sources

• OpenAI targets work of Wall Street junior bankers with new ChatGPT for Financial Services
• More capable AI is more dangerous AI, Anthropic learns
• Newsom signs laws aimed at protecting kids from risks of social media, AI chatbots
• California breaks ground with first AI and social media laws curbing teenage use in the nation
• An ex-Anthropic researcher claims AI could kill us all by 2030. But he fails to answer the most essential question: What are we supposed to do about it?
• US listing could put Japan's Kioxia in global AI spotlight, Voya's Thomas says
• The Only 2 Moats That Actually Work In The AI Era
• As banks' AI use evolves, core truths about the business still apply
• BigBear.ai Stock Dives 29% in the Past 3 Months: Hold or Fold?
• Google completes deal for AI agents startup Mechanize
• Nvidia CEO Reportedly Dismisses Anthropic Exec's AI Doomsday Warning As ‘Outlandish’ And ‘Deeply Untrue’