AI Evaluations Look Unbiased Without Being Fair
New research by Prof. Tristan Botelho and PhD student Qingyang Wang finds that LLMs avoid overt discrimination when evaluating candidates. They still rely on racial and gender cues.
Read the original →Article digest · Oct 9 PM
AI Ethics — what each story says
Every story in this issue with a short summary, so you can decide what to read before leaving the site. Summaries are written by AI (Claude) only from each publisher's own description and checked by Jev for anything not in that description. They are not full translations. Follow "Read the original" for the full story.
27 stories19 with summaries8 topics
⚖️ Bias, discrimination and fairness3
- Headlines only (the publisher did not provide a description)
- ‘Artificial Intelligence Is Not Gender-Neutral,’ Third Committee Hears, as Experts Warn of Digital Violence, Backlash against Women’s RightsUN Meetings Coverage and Press Releases · 10/08 04:00 UTC
- New Hampshire Board of Pharmacy Proposes Rules Governing Artificial Intelligence in Pharmacy PracticeQuarles & Brady · 10/09 00:59 UTC
🔒 Privacy and surveillance2
Anthropic bans ‘abusive or cruel behavior’ toward Claude
The usage policy update also adds new rules addressing propaganda campaigns, surveillance and weapon development.
Read the original →Many Quebec voters got false election information from chatbot, MPs hear
A Parliamentary committee heard that many Canadians who turn to chatbots for election information do not always get accurate responses. McGill University academics have been tracking chatbot activity, including during the recent Quebec election. McGill political scientist Aengus Bridgman said an estimated 40 per cent of Quebec voters
Read the original →🔍 Transparency and accountability4
Lawmakers eye more federal action on AI standards, security, disclosures
Lawmakers appear keen for agencies such as CISA and NIST to take more leading roles in evaluating AI security risks and setting safety standards.
Read the original →Can your AI compliance decisions survive a regulator’s audit?
Financial institutions are relying more heavily on AI to support compliance decisions, which raises the question of whether they can prove how those decisions were made. According to Duna, many firms may struggle to answer convincingly when supervisors review them. Supervisors reviewing an AI-assisted decision expect to see a clear record of how it was reached.
Read the original →In the face of AI, auditors need to know what makes them unique
Richard Chambers, the former head of the Institute of Internal Auditors, has written his fifth book, 'Trusted Advisors: The Human Edge in the World of AI.'
Read the original →- Headlines only (the publisher did not provide a description)
- Making sense of AI on social mediaThe Express Tribune · 10/09 00:13 UTC
🎨 Creators, copyright and consent1
- Headlines only (the publisher did not provide a description)
- VAYLA Arena and the AI Music Discovery Challenge — Why Human Participation Matters in an Age of Expanding CreationStreetInsider · 10/09 09:06 UTC
🎭 Misinformation and deepfakes5
AI-generated videos get labels. The European Union is applying the same standards to text.
The European Union wants LLMs to include labels that identify AI-generated text. ChatGPT will be the first to be subject to this.
Read the original →We're making it easier to identify AI-generated content globally.
Google is launching a standalone platform to help people identify whether online content was created using Google AI or tools from its industry partners.
Read the original →SynthID: How Google’s AI Watermark Works and Who’s Participating
SynthID marks AI-generated content in a way that is invisible to humans but detectable by software. The article explains how the technology works and lists partners including OpenAI, Nvidia and Apple.
Read the original →- Headlines only (the publisher did not provide a description)
- Google opens AI content detector to the public as synthetic media growsBusinessday NG · 10/08 18:57 UTC
- Google's tool for detecting AI-generated content is now available to all usersMezha · 10/07 15:50 UTC
🧒 Protecting children and vulnerable people2
Helping kids question AI: interactive voice-based chatbot explores trust, accuracy and advice
As children increasingly turn to AI for information, advice and support, researchers are exploring how the way chatbots respond can shape young people's understanding of these technologies.
Read the original →- Headlines only (the publisher did not provide a description)
- Health West: Teens using AI for mental health: Benefits and risksIdaho State Journal · 10/08 16:25 UTC
📜 Ethics principles and guidelines5
EY revenue hits USD $57 billion as AI demand surges
AI-related work helped push EY's annual revenue to USD $57 billion. Clients are shifting from trials to wider business use.
Read the original →AI Review Committee provides expert advice on governance
An AI Review Committee has been established to provide governance and oversight for public sector AI use across Australian government agencies.
Read the original →Singapore sets AI risk rules for financial sector from 2027
On Wednesday, the Monetary Authority of Singapore (MAS) issued the Guidelines on Artificial Intelligence (AI) Risk Management to support responsible AI adoption.
Read the original →Students build AI-powered solutions for education at Minds and Machines Hackathon
The University of Utah One-U Responsible AI Initiative's Minds and Machines: AI in Education Hackathon ran over 48 hours in mid-September. It brought together 102 participants.
Read the original →- Headlines only (the publisher did not provide a description)
- Opinion: ‘Responsible AI’ will be proven on Chicago’s factory floorsCrain's Chicago Business · 10/08 20:41 UTC
🇯🇵 AI ethics in Japan5
Copyright Wars, Part 5: AI spreads as stakeholders seek ways to build a system
Songs created with generative AI are increasing rapidly. French music streaming service Deezer announced in July that about 90,000 AI-generated tracks are uploaded per day. It said many of them have improper purposes.
Read the original →"I don't want to show kids generative AI videos": Reiwa Roman's Kemuri Matsui on parenting
Comedian Kemuri Matsui of Reiwa Roman and Mai Tsujimoto, a mother of three, discuss how far parents should watch over and where they should guide children's interests. They talk about lessons, games and YouTube, and how to nurture children's free thinking.
Read the original →Sony Music requests removal of over 260,000 AI deepfake songs, double the figure from six months earlier
Sony Music has said it asked digital platforms to take down more than 260,000 AI-generated deepfake songs imitating its artists. According to the Financial Times, the number of takedown requests nearly doubled over the past six months, as of the end of September 2026.
Read the original →Can technological progress and IP protection coexist? New rules for generative AI "learning"
The article summarizes in Q&A format the new rules for generative AI.
Read the original →How Amazon Bedrock Automated Reasoning Checks Change Generative AI Compliance
Probabilistic checks such as LLM-as-a-judge, where another large language model evaluates output, cannot give regulated industries the formal assurance that audits require. The article explains how Automated Reasoning checks in Amazon Bedrock Guardrails use formal verification to mathematically verify AI output, with evidence ready for audits. It also covers adoption cases from six industries, including Amazon Logistics, Lucid Motors, Fortive and FETG, a reference architecture for compliance checks, and getting-started steps.
Read the original →The analysis and visual summary are in the Oct 9 PM issue. Japanese version here.