AI Ethics News — 2026-11-09 (Mon) Daily News

Overview

The past 48 hours brought several ethical questions into view: AI systems used to monitor students without their agreement, with a Chinese university dropping such a feature after an outcry (South China Morning Post, WION). Studies and reports also examined fairness in how AI models treat English accents and how LLM recommenders amplify bias (Qazinform, Chosunbiz). Separately, AI-generated images were reported in extortion scams and in a claim about a Nigerian politician's health (Bitdefender, FactCheckHub). 【Bias, discrimination and fairness】 - A report says AI models rank English accents differently, raising fairness concerns (Qazinform). The headline does not say which accents are disadvantaged, or how the ranking was measured or at what scale. - A tool called EchoTrace reveals that LLM recommenders amplify bias and errors through feedback loops (Chosunbiz). - A hybrid AI framework that evaluates teaching effectiveness in higher education is described as reducing bias (Bioengineer.org). This is the framework's stated aim; the headline gives no evidence of how much bias was reduced.

Privacy and surveillance

- A Chinese university stopped using AI to spot absent-minded students after an outcry (South China Morning Post). Another report describes the same case as the university scrapping an AI feature that tracked students' attentiveness after a privacy backlash (WION). - The headlines show the feature measured student attentiveness and that students and others objected. They do not say what data was collected, whether students consented, or how the data was stored. 【Transparency and accountability】 - A systematic review evaluates explainable AI and the RDoC framework in Alzheimer's diagnosis (Bioengineer.org). - Britannica asks how human oversight can make AI systems more accountable (Britannica). - Scientists and whistleblowers give grim predictions about AI, including a "team of hackers that never sleeps" (Nine.com.au). The headline gives no detail on what is disclosed or withheld. 【Creators, copyright and consent】 - There are no notable developments in the headlines for this section.

Disinformation and deepfakes

- Scammers use AI-generated images of missing pets to extort desperate owners (Bitdefender). This is reported as a scam practice that harms pet owners. - AI-generated watermarks are becoming "spy marks" used to track users (GIGAZINE). This is the outlet's framing of a tracking concern. - An image of Peter Obi on a sick bed was found to be AI-generated (FactCheckHub). The fact-check outlet reports it as confirmed AI-generated. - A workshop in Victoria teaches residents to spot AI-generated misinformation (crossroadstoday.com). This is a community-level response. 【Protecting children and vulnerable people】 - An opinion piece argues that AI is not going away and that parents need to teach children how to use it (Sun Sentinel). This is an opinion, not a policy decision. - A critical review examines how AI is conceptualized in early childhood education (Bioengineer.org). - Desperate pet owners are the targets of the extortion scams noted above (Bitdefender). 【Ethical principles and guidelines】 - Students voted in favour of Ethical AI and of disaffiliating from NUS, according to a breaking report (The Boar). The headline does not say what the Ethical AI vote entails. - Experts at IIT Ropar advocate compassionate and responsible AI development (EdexLive). This is advocacy, not an adopted rule. - A Türkiye minister warns that AI responsibility cannot be delegated (Yeni Şafak English). This is a statement, not a decision.

Japan's AI ethics

- Masahiro Sakurai commented on generative AI, saying it "must not become exploitation," and also voiced concern (Niconico News). Per the headline, the concern is about exploitation, though the specific parties involved are not named.

Statements by key figures

No notable developments. None of the tracked figures (Timnit Gebru, Fei-Fei Li, Yoshua Bengio) appear in the headlines. 【Integrated analysis: what is happening】 - Lack of consent and notice appears in the student attentiveness case, where the backlash suggests people objected to being monitored (South China Morning Post, WION), and in the watermark tracking concern (GIGAZINE). This may indicate that tracking tools are drawing scrutiny when users have little say. - Lack of explanation appears in the English accent rankings and the LLM recommender feedback loops (Qazinform, Chosunbiz), where affected people may not know how they were judged. The attention given to human oversight and explainable AI (Britannica, Bioengineer.org) suggests demand for ways to check such systems. - Harm from AI-generated content falls on people with little protection, such as pet owners and the public who see fake images (Bitdefender, FactCheckHub). Responses so far appear to rely on education and fact-checking (crossroadstoday.com, FactCheckHub). - For companies and institutions, the university case suggests that features monitoring individuals may be withdrawn if users object. Organizations using AI to evaluate people may need to be ready to show how bias was measured.

What to watch

- Whether the Chinese university's withdrawal prompts other institutions to review similar attentiveness-tracking features. This is an inference. - Whether the accent-ranking and recommender-bias findings lead to disclosed testing methods or fixes by the developers involved. - Whether the NUS Ethical AI vote and the calls for responsible AI from IIT Ropar and the Türkiye minister turn into concrete rules or commitments.