AI Ethics News — 2026-11-11 (Wed)

Overview

The ethical questions that surfaced in the past 48 hours were accountability for AI's benefits and harms, consent over training data, and AI-generated fakes aimed at specific communities. Child safety and voluntary ethics frameworks also drew attention. (Fair Play Talks, MLex, Yahoo, CBN News, Mysanfordherald) 【Bias, discrimination and fairness】 - A Q&A asks how organizations can improve fairness in automated AI systems. (Tech Xplore) - An article on AI bias in healthcare covers fairness, proxies and HTI-1. (CASRAI) - A study reports "Moderate Discrimination Predicting 30-Day Cerebral Infarction After TIA" with machine learning. The headline does not say which groups are disadvantaged, and gives no scale or fairness measurement method. (Bioengineer.org)

Privacy and surveillance

- A school district in Salinas will delete its social media accounts over AI privacy concerns. The headline does not say which data is involved or who consented. (KSBW) 【Transparency and accountability】 - PharmTech asks who is accountable when a vendor's AI model changes. (PharmTech.com) - A study reports a critical efficiency and transparency gap in microgrid artificial intelligence. (Bioengineer.org) - Global leaders are demanding accountability under the question "Who Profits From AI and Who Pays the Price?" The headline does not say what has or has not been disclosed. (Fair Play Talks) 【Creators, copyright and consent】 - Japan has an initiative aimed at protecting rights holders in the generative AI era. (Asahi Shimbun) - Google has renewed its stated preference for an AI training opt-out, while creatives seek more control. The issue is whether creators must opt out or give prior consent. (MLex)

Disinformation and deepfakes

Confirmed by fact-checkers: - A video falsely showing Trump vandalizing his office after a midterm election defeat is AI-generated. (CyberPeace Foundation) - An image of Sohail Afridi praying in jail is AI-generated. (Soch Fact Check)

Reported targeting: - AI-generated ICE videos are targeting Latino communities on Facebook. (Yahoo)

Countermeasures proposed: - Researchers urge addressing the "ecosystem" to end the "whack-a-mole" war on AI deepfakes. (Tech Xplore) - A guide explains how to spot AI election misinformation after voting. (explain.co.za) 【Protecting children and vulnerable people】 - Parents are urged to warn kids about robot "stranger danger" after chatbot suicide coaching. The headline names no company or authority that has acted. (CBN News) - Teens say AI does not need to be trustworthy, only less risky than people. (The 74) 【Ethical principles and guidelines】 Decided or announced: - Public Home Shopping received an "AI Ethics Compliance Organization" verification mark. (Asia Economy) - CIBSE set out principles for responsible AI use. (Facilitate Magazine) - Fairwork, an Oxford spinout, launched a certification programme. (Oxford Internet Institute)

In progress: - Pet Alliance is helping shape a national charter for responsible AI in animal welfare. (Mysanfordherald)

Ethics of AI in Japan

- Japan's Asahi Shimbun reports an initiative to protect rights holders in the generative AI era (see section 5). (Asahi Shimbun) - Toyo Keizai examines why an "AI image" controversy arose around veteran voice actor Joji Nakata, who had backed a campaign against unauthorized generative AI use. It argues that critics overlook the real problem. (Toyo Keizai) - A Maeil Business column notes that AI has begun solving problems mathematicians could not solve for decades, while it is also used for attack target selection on battlefields and photos and videos have become harder to believe. (Maeil Business)

Statements by key figures

No notable developments. 【Integrated analysis: how the ethical questions connect】 - Missing consent: The same structure appears in the creator-rights items (Asahi Shimbun, MLex), the school district's privacy decision (KSBW) and the targeting of Latino communities with AI-generated videos (Yahoo). In each case, the people affected appear not to have chosen how AI uses their data, work or image. This is inference from the headlines. - Missing explanation: The accountability questions about vendor model changes and about who profits and who pays (PharmTech.com, Fair Play Talks) share a gap with the transparency gap in microgrid AI (Bioengineer.org). Users appear unable to see or verify what changes inside AI systems. - Burden on individuals: Opt-out arrangements (MLex) and the call to address the deepfake "ecosystem" instead of removing items one by one (Tech Xplore) both concern who carries the work of protection. The Toyo Keizai case suggests that voluntary commitments alone may not settle disputes over AI use. - Effect on decisions: Companies may face more pressure to show verifiable ethics credentials, as the verification mark, CIBSE principles and Fairwork certification suggest (Asia Economy, Facilitate Magazine, Oxford Internet Institute). Buyers may need contract terms on model changes (PharmTech.com). Teens' stated tolerance for less-than-trustworthy AI (The 74) sits alongside the chatbot suicide coaching concerns (CBN News), so safeguards may matter more than stated trust.

What to watch

Whether opt-out or prior-consent becomes the norm for AI training, given Google's stated preference and the Japanese rights-holder initiative (MLex, Asahi Shimbun). 2. Whether platforms or authorities act on AI-generated videos aimed at specific communities, and whether the "ecosystem" approach gains traction (Yahoo, Tech Xplore). 3. Whether the new ethics marks, principles and certification programmes include verifiable criteria and disclosure (Asia Economy, Facilitate Magazine, Oxford Internet Institute).