AI Ethics News — 2026-10-26 (Mon) Daily News

Overview

The ethical questions that surfaced in the past 48 hours were discrimination in AI-assisted policing, new forms of gender-based violence, and labeling of AI-generated content in politics and on social media. Deepfake-style deception was also in the news, in a scam and a political attack ad. In Japan, personal data of employees and family members entered a generative AI tool that the company had banned. 【Bias, discrimination and fairness】 - Research warns that AI in policing risks deepening discrimination. The headline does not give the measurement method, the scale or the groups affected (BBC). - The UN says AI has given rise to new and dangerous forms of gender-based violence. The headline does not say how this was measured (logos-pres.md). - A letter to the editor urges a delay in deploying artificial intelligence (Grand Forks Herald). An article argues that the first job of AI-era leadership is protecting judgment (The European Business Review).

Privacy and surveillance

- Personal information of employees and their families, covering about 334,000 people across 4 categories, was reported in connection with a generative AI that Panasonic had banned from use. The headline does not say whether the people consented (innovaTopia). See section 9. 【Transparency and accountability】 - Taiwan's MODA is working on better disclosure of AI on social media (Taipei Times). The headline does not say what will be disclosed or when. - An article on healthcare AI frames the issue as moving "From wide adoption to operational accountability" (RSM US LLP). - The headlines do not show what AI developers or deployers have disclosed. 【Creators, copyright and consent】 - A Substack post asks "Will Generative AI do more harm than good?" The headline names no specific parties or copyright dispute (Marcus on AI | Substack).

Disinformation and deepfakes

- Confirmed as reported: Eaton County Sheriff warns of a new scam that uses AI-generated photos of deputies and demands Bitcoin (WILX). - Claim: an Asheville drag queen says an attack ad that falsely links her to Jamie Ager was made with AI. The headline reports her statement, not an independent finding (WLOS). - Countermeasures: Shontel Brown commented on labeling AI-generated political ads (Modern Ghana). India's NCERT introduced a 20-hour AI module for classes 9-12 covering ethical boundaries to detecting deepfakes (ET Enterprise AI). 【Protecting children and vulnerable people】 - The UN warns of new forms of gender-based violence linked to AI (logos-pres.md). - India's NCERT introduced the deepfake-related module for students in classes 9-12 (ET Enterprise AI). - The drag queen in Asheville says she was falsely linked to a political figure through an AI-made ad (WLOS). 【Ethical principles and guidelines】 - Decided: NCERT introduced a 20-hour AI module for classes 9-12 (ET Enterprise AI). - In progress: MODA is working on better social media AI disclosure (Taipei Times). - Positions and proposals: Shontel Brown on labeling AI-generated political ads (Modern Ghana). A letter calls for delaying AI deployment (Grand Forks Herald).

Japan's AI ethics

- Panasonic: personal information of employees and family members, about 334,000 people across 4 categories, was reported in connection with a generative AI banned from use (innovaTopia). - au Web Portal published a Q&A on generative AI usage risks to be aware of, as part of its "Digital & AI Gear" complete guide (au Webポータル).

Statements by key figures

No notable developments. 【Integrated analysis: how the ethical questions connect】 - Lack of consent appears in several items. Personal data of about 334,000 people in a banned generative AI tool (innovaTopia), AI-generated photos of deputies used in a scam (WILX), and an AI-made ad that falsely links a person to a political figure (WLOS) all involve people whose data or likeness was used. The headlines do not show that any of them agreed, and this is an inference. - Lack of disclosure is the second shared structure. MODA's disclosure work (Taipei Times), the labeling of political ads (Modern Ghana) and the call for operational accountability in healthcare (RSM US LLP) all address the question of who is told that AI was involved. - Likely effects on decisions: companies may need to check whether banned tools are actually blocked and what data staff enter into them. Platforms and campaigns may face labeling expectations. Users may need to verify unfamiliar images and ads, which is the skill the NCERT module targets (ET Enterprise AI). These are inferences.

What to watch

Whether MODA's social media AI disclosure work leads to a concrete rule, and what it requires (Taipei Times). 2. Whether labeling of AI-generated political ads moves from statements to adopted requirements (Modern Ghana). This may be tested by further disputed ads like the one in Asheville (WLOS). 3. Whether more detail emerges on the Panasonic case, including how the data reached the banned generative AI and what is done about it (innovaTopia).