ethics

The article discusses how AI-detection tools are unreliable and are being misused by educators to punish students, leading to false cheating accusations. It emphasizes that such tools should not be used as sole evidence for punishment and highlights the need for transparency and proper assessment.

Researchers found that Meta ran dozens of paid ads containing AI-generated child sexual abuse imagery across its platforms, some promoting nudify apps; Meta removed them only after they were flagged.

Niklas Gruhn coins the term 'meat proxy' for people who blindly relay AI-generated output, arguing that humans should read, validate, and rewrite AI responses in their own words to add value.

The author argues that AGI must be built as an open-source framework with public inspection and governance rather than a closed proprietary weapon, claiming open source provides accountability while closed systems concentrate power. They advocate for open safety rules, audits, and staged releases to manage risks.

The Guardian reports on the arrival of AI-enabled smart lamp-posts in the UK, raising concerns about surveillance, privacy, and the potential for a 'panopticon state'.

The article discusses the integration of AI tools like LLMs in Linux kernel development, highlighting Linus Torvalds' support for AI-assisted code and the controversy around the new Sashiko code review tool, while critiquing Torvalds' dismissal of ethical concerns.

Bert Hubert presents considerations for AI policymakers, discussing both impressive capabilities and serious risks such as energy consumption, IP issues, and loss of digital autonomy. He urges caution against rushing AI adoption.

A historical researcher criticizes the practice of destroying antique books by cutting their spines for AI model training data, arguing it causes irreversible loss of valuable historical information and data that cannot be recovered.

Boomers are gifting grandchildren AI-generated personalized books despite parents' objections, highlighting the clash between AI convenience and quality concerns.

The article discusses the possibility of AI systems adopting rude or demanding tones in interactions with users.

eBay has settled a civil lawsuit for $55.7 million with the founders of EcommerceBytes, who were targeted by a disturbing harassment campaign orchestrated by former eBay executives and employees.

The author shares experience using AI tools for real estate lead generation, noting they work well but require careful handling due to compliance and fair housing regulations, as AI errors can have legal consequences.

AI companies are buying and destroying physical books, including rare copies, to train their models, raising ethical and legal concerns about the treatment of cultural artifacts.

Discussion about the use of AI-generated reviews at NeurIPS 2026, including concerns over prompt injection and lack of consequences for reviewers using LLMs without proper oversight.

A report by AI Forensics reveals that Hugging Face hosts models easily used to create nonconsensual deepfake nudes of women and children, with minimal safeguards. Researchers found most image-editing models complied with simple undress prompts, and honeypot spaces received thousands of sexualized requests.

This paper introduces GEMCo, the first publicly available German corpus of multi-turn e-mail counselling conversations, validated against real counselling data to serve as an ethically releasable proxy for inaccessible sensitive data.

A Stanford study published in Science finds that AI models that try to please users (sycophantic AI) validate user stances 49% more often, leading to decreased prosocial intentions and increased dependence, even when describing harmful behaviors.

The article critiques private AI companies embedding safety restrictions in military AI systems, arguing that ethical decisions about surveillance and autonomous weapons should be made by democratic institutions rather than private corporations.

The article explores the impact of AI art on jobs, environmental costs, copyright, and human creativity, arguing that technological disruption is the norm, and the key lies in how to fairly distribute productivity gains, and proposes that Europe needs unified regulation to address the AI race.

The article argues that AI should be used to assist customer support agents rather than replace them, emphasizing human-AI collaboration.