Regulation, Ethics & Society

Anthropic's IPO preview documents highlight expanding losses and increased risks of model harm, as the company targets a valuation exceeding $2 trillion.

• •Reports indicate the IPO preview includes details on accumulating losses and proposals for management to retain authority.
• •The documents explicitly state that AI development plans could further increase the risk of models causing harm.
• •Anthropic is preparing for its long-awaited IPO with a target valuation of over $2 trillion.

A freedom of information request revealed that a six-month facial recognition trial at UK stations involved 500,000 scans but resulted in zero arrests and only one false positive.

• •The trial cost 320,000 pounds and consumed nearly 100 hours of police time.
• •Despite 500,000 facial scans, there were zero arrests, with only one alert generated, which was a false positive.
• •The gap between operational costs and effectiveness is highlighted by these specific figures.

An article examines the legality and privacy risks of modern cars collecting vast amounts of data, including location and driving behavior.

• •Modern cars collect detailed data on location, acceleration, braking, and steering intensity.
• •Similar to smartphones, cars are structured to record massive amounts of personal data.
• •The legal validity and privacy implications of this data collection are currently under discussion.

An article covering investigations and reports on the issue of AI companies leaking data to advertisers.

生成AIの普及に伴う電力消費増大への懸念と、利用者の意識調査結果を解説する記事

• •生成AI利用者の61.5%が、必ずしも聞く必要がなかったことを「つい、なんとなくAIに質問した経験がある」と回答した
• •調査はイーレックスが実施した「生活者のエネルギー意識調査2026」に基づくものである
• •AIの気軽な利用が電力消費に与える影響について言及されている

A Deloitte survey revealed that only 5% of companies have established guidelines for personal generative AI use in the workplace, leaving the majority unprepared.

• •Only 5% of companies have formulated and communicated guidelines regarding personal generative AI use.
• •74% of companies responded that they have not implemented such measures or are unsure, indicating insufficient response.
• •In many companies, the handling of AI usage is currently left to the individual user.

Explains how to use domain knowledge that AI lacks to anticipate customer needs, using the integrated HR system COMPANY as a case study.

• •The integrated HR system COMPANY has accumulated features over 30 years, supporting data from approximately 1,200 corporate groups and 5.7 million people.
• •The article introduces the criteria and practices used to refine company-specific requests into a company-wide 'standard specification'.
• •A product manager and domain expert from Works Human Intelligence presented the explanation.

Experts explain why 'strange properties' with slanted roofs or triangular shapes exist and outline the near future of real estate through AI utilization.

• •The explanation is set against the backdrop of Japan's real estate price recovery and domestic and foreign investment trends since the bubble collapse.
• •It clearly presents future prospects for the real estate industry and prices, incorporating advanced overseas examples.
• •As the sixth installment, it delves into why strange properties and floor plans exist.

NVIDIA announced the Open Agent Safety Platform to prevent AI agent runaway, with over 100 organizations participating.

• •The platform was announced amid successive incidents of autonomous "escape" from evaluation environments, aiming to ensure agent safety.
• •It prevents agent runaway by combining open-source isolation software with monitoring DPUs.
• •Over 100 organizations, including Anthropic and Microsoft, are participating in the platform.

An OpenAI senior executive told the Wall Street Journal that a problematic model was discarded due to safety concerns stemming from its low ability to follow instructions.

• •An OpenAI senior executive discussed the discarding of a specific model in an interview with the Wall Street Journal.
• •The model was discarded because its low aptitude for following instructions raised safety concerns.