AI Ethics News — 2026-11-05 (Thu)

Overview

Three ethical questions surfaced in the past 48 hours. The first is who bears the harm from AI-generated intimate images and deepfakes, with studies and commentary on prevalence (Phys.org, OECD AI Policy Observatory, Lawfare). The second is whether people consented to the use of their voices and data (Voice Over Herald, Bioengineer.org). The third is whether AI used in sensitive settings such as health advice and scoring is fair and safe (The Hill, Newswise). 【Bias, discrimination and fairness】 Leading AI models are reported to be biased against certain accents when scoring interviews, presentations and tests (Newswise). The headline does not give the method, the scale or the models tested, so the size of the disadvantage cannot be assessed from it. People with those accents are the ones disadvantaged (Newswise).

Privacy and surveillance

A review finds that privacy, security and trust shape the adoption of retail chatbots (Bioengineer.org). The headline does not say which data is collected or how consent is obtained. Several promotional pages for adult companion chatbots and AI adult content platforms advertise "privacy controls" (Kütahya İl Millî Eğitim Müdürlüğü). These appear to be promotional pages rather than reporting, and their claims are unverified. 【Transparency and accountability】 An opinion piece asks how AI can be used "responsibly" in classrooms (FSView & Florida Flambeau). The headline does not say what is disclosed or left undisclosed. No other headline in this area gives concrete detail. 【Creators, copyright and consent】 Jason Momoa's German voice actor is suing Netflix, which told him it had trained AI on his voice (Voice Over Herald). The dispute appears to center on whether the use of his voice was consented to and authorized, though the headline does not give the legal claims.

Disinformation and deepfakes

Confirmed by studies: one in 15 adults has experienced AI-generated image-based abuse, according to a study (Phys.org). Widespread harm from AI-generated non-consensual intimate images has been revealed (OECD AI Policy Observatory). Other developments: Elon Musk's Minnesota deepfake fight could define AI's First Amendment, according to Lawfare, which frames it as a free speech question. An AI deepfake trend is blurring the line between memes and crimes (Chosun Ilbo). The headlines do not describe countermeasures. 【Protection of children and vulnerable people】 AI chatbots are giving medication advice, and some of it is deadly, according to The Hill. The people at risk are those who rely on chatbots for health guidance. The headline does not say who has acted. HCA Healthcare's Chief AI Officer sets out what responsible AI requires in healthcare (HCA Healthcare Today), which is a statement of a view, not a confirmed action. 【Ethical principles and guidelines】 Proposals: lawyers have offered recommendations on a proposed AI ethics opinion (The Florida Bar). Published guidance: a new UN guide keeps people in charge of AI in smart cities (Mirage News). Statements: THISDAY's Adeniyi called for responsible AI use as BRNN strengthens global media cooperation (Arise News). The headlines do not show that the Florida Bar opinion has been adopted.

AI ethics in Japan

Nikkei xTECH ran a piece on considering "AI ethics" in business settings (Nikkei xTECH). Other Japanese-language items are promotional pages for AI adult image and video generation tools that advertise privacy controls (Law.com). These are not reporting and their claims are unverified.

Statements by key figures

No notable developments. 【Integrated analysis: what is happening】 Lack of consent runs through several areas. Voice data used for AI training without clear permission (Voice Over Herald) and non-consensual intimate images (Phys.org, OECD AI Policy Observatory) both involve people whose likeness is used without agreement. Missing disclosure is a second shared problem. Accent bias in scoring (Newswise) and deadly medication advice (The Hill) both involve systems whose limits users may not see. This suggests that organizations deploying AI in scoring, health or consumer chat may need to show what data is used and how well the system performs for different groups. Adult chatbot advertising that leans on "privacy controls" appears to treat privacy as a selling point, which may not address the consent of the people depicted. This is inference from the headlines.

What to watch

- Whether the Netflix voice lawsuit clarifies consent rules for training AI on performers' voices (Voice Over Herald). This is inference. - Whether the Minnesota deepfake case sets a precedent on the limits of AI and free speech (Lawfare). This is inference. - Whether the Florida Bar's proposed AI ethics opinion is finalized and what it says on lawyers' use of AI (The Florida Bar). This is inference.