AI Daily News — 2026-10-10 (Sat)

AI news analysis report (past 12 hours)

Note: The following analysis is based on headlines only; the text of the article has not been verified. Avoid making assertions that go beyond what can be gleaned from the headline, and clearly indicate that speculation is speculation.

Overview

Anthropic has announced a policy prohibiting "abuse and cruelty" towards Claude, sparking a debate over the welfare of AI models. OpenAI has had controversial issues with its annual sales forecast of $70 billion and the firing of three safety researchers. Google has launched Gemini Agent for businesses, and the axis of competition is shifting from "models" to "agents."

Anthropic / Claude trends

The BBC and CBS reported that Anthropic's terms of service prohibit any "abusive or cruel" behavior towards Claude. Business Insider has compiled a list of experts' reactions to the policy. Meanwhile, free startup benefits were suspended within days of launch, according to Forbes. It appears that there was some kind of problem with demand or operations, but the reason is unclear from the headline. ABC reported Anthropic's announcement that it had shut down potential use as a bioweapon. 6abc reports that the Philadelphia Police Department has flagged a false report in an unsolved murder case attributed to Claude, which is a serious example of Hallucination's actual harm. SCMP is verifying Anthropic's claims that Chinese AI companies are secretly using Claude.

OpenAI trends

Bloomberg reported that OpenAI expects annual sales of $70 billion by the end of 2026. The NYT reports on the aftermath of the "revenue surprise," but the direction (upward or downward) cannot be determined from the headline. According to ABC, three safety researchers have been fired. OpenAI explained to Business Insider that it was not because of what it said about the risks. Mathematicians strongly oppose the publication of hundreds of AI-generated proofs (Futurism). Accelerated growth and friction with the safety and academic communities are simultaneously surfacing.

Google AI trends

Google has announced an integrated Gemini Agent for businesses that can work across multiple apps. Futurum described it as a strategy that ``puts agents before models,'' and Yahoo Finance positioned it as a strategy to lock in corporate AI infrastructure. ZDNET's examination of Gemini Spark suggests that its valuation is likely to be diluted. It was also reported that Gemini had partially malfunctioned in various parts of the world, causing anxiety at a time when expectations were high for its availability as it became an agent.

Trends of other AI companies

- Microsoft: Announces Surface Laptop Ultra and local AI agent to turn Dynamics 365 into an "ambient CRM." The policy to continue Copilot+ PC has been criticized, and some are considering canceling Microsoft 365 Family. - xAI / Tesla: Tesla is reportedly encouraging employees to use Grok 4.5 under AI spending caps. There is also a plan for Grok to allocate users to other companies' models. - Mistral: Released Large 4 and strengthened the open model camp in the West. - Chinese: According to Reuters, Chinese AI developers publish safety tests in only 3.6% of model releases. Naive AI, run by a professor at Tsinghua University, has reached a valuation of $1.4 billion. - Meta: NYT reported how Zuckerberg decided to introduce AI agents.

Statements and trends of AI celebrities

- Sam Altman (OpenAI CEO): Vanity Fair featured an article by the family of ChatGPT and their daughter's suicide. Fortune reported that he said, ``We need a new renaissance.'' Others have described Musk as "bully." - Elon Musk(xAI / Tesla): Warned that "compute crunch is coming, people don't understand." He mocked SoftBank's plan to raise $100 billion, saying, "Only a fool would pay for it." - Jensen Huang (NVIDIA CEO): Received America's highest scientific honor (TechSpot). It was also reported that he and Lisa Su (AMD CEO) agreed on one point. - Dario Amodei (Anthropic CEO): Axios reported that AI companies are developing scenarios after a "public backlash." Vox argues that AI leaders' fears may be genuine. - Sundar Pichai (Google CEO): Announced Gemini Agent. There is also a story about how my father spent his annual salary on airline tickets at Stanford. - Mark Zuckerberg (Meta CEO): Explained to investors that AI investments are accelerating all businesses. He also said that AI should be regulated globally. - Demis Hassabis (DeepMind CEO): Related Isomorphic Labs was reported to be aiming for a valuation of $40 billion (May's round was listed as $2.1 billion). - Ilya Sutskever, Yann LeCun, Geoffrey Hinton: New statements by the authors are not included in the headlines and are only linked to related articles. - Jeff Bezos (not on tracking list): Talks about the arrival of a three-day work week, sparking backlash.

Breakthrough and innovation in AI technology

AI's presence in mathematics is rapidly increasing, with The Times even describing it as a "hostile takeover." On the other hand, there is also strong opposition regarding the quality of proof and verification. In the medical field, Johns Hopkins' use of AI in cancer research and Utah's trial of AI-based prescriptions are progressing. OpenAI's Daybreak reduced Sophos' threat investigation time by 96% (case claim). GlobalFoundries announces next-generation platform for Physical AI.

AI Regulation/Policy/Social Impact

On the employment front, Amazon continues to cut jobs while denying any connection to AI, and HSBC reportedly plans to cut up to 70% of its wealth advisor jobs in the UK. In Silicon Valley, even high-income earners are finding it difficult to find work again. In the legal field, California's Generated AI Rules for Lawyers were put in place, and gaps in criminal liability for harm caused by AI agents were also noted. The UK ICO has received commitments from 10 companies to handle AI data. Flock Safety has cut staff amid public opposition to AI cameras.

Integrated analysis: What is happening now

The axis of competition has shifted from model performance to ``actual operation of agents.'' Google, Microsoft, Meta, and Tesla have launched cross-functional agents. Although OpenAI leads in terms of revenue, it has a credibility liability due to the firing of its security staff and opposition from the academic community. While Anthropic sought to differentiate itself in terms of ethics such as model welfare and prevention of abuse, it also showed instability in its operations. On the capital front, the concentration of funds and computing resources will continue, as evidenced by the compute crunch mentioned by Musk, SoftBank's huge fundraising, and Isomorphic Labs' rising valuation. Geopolitically, the axes of conflict are the lack of safety disclosure by Chinese companies, Anthropic's suspicions that Chinese companies are using Claude, and the rise of Western open models such as Mistral. Uncertainty about employment has created fears of a ``public backlash,'' and companies are moving to provide negative explanations and prepare scenarios in advance.

Future attention points

The background to the dismissal of OpenAI's safety researchers and the impact of sales forecasts on the market. 2. Corporate implementation results of Gemini Agent, Microsoft, and Meta agents, and responses to failures and misinformation. 3. Public opinion and regulatory movements against AI-induced job cuts, and tight compute supply.