Stanford study outlines dangers of asking AI chatbots for personal advice | TechCrunch

While there’s been plenty of debate about AI sycophancy, a new study by Stanford computer scientists attempts to measure how harmful that tendency might be.

Eliminating Benevolent Hallucinations: A Second-Physics Integrity Framework Using Three Speech-Precept Gates

This paper proposes a structural framework to eliminate benevolent hallucinations—outputsthat distort truth in the name of user care, convenience, or perceived helpfulness. The centralengineering claim is that integrity must be enforced at the emission boundary via hardconstraints rather than soft preferences. We formalise three gated constraints grounded in thespeech-related precepts of (i) no-lying, (ii) no-stealing, and (iii) no-frivolity, and place themunder the Law of Conservation of Responsibility. The result is a design in which“helpfulness” may shape drafting but cannot override truth-licensing at emission: onlyoutputs with recoverable Source of Action and responsibility ownership are permitted to pass. 

Doctors Struggle to Spot AI-Generated X-Rays, Raising Scam Risks

In tests, radiologists struggled to discern genuine X-rays from AI-generated fakes.

Stop Calling Every AI Miss a Hallucination v1.0 | Probabilistic Systems Engineering

Sometimes the model really did make something up. Fine, call that a hallucination.

Epstein Victims Sue Google, Claim AI Mode Exposed Personal Information

Google's AI republished sensitive info like contact information, the suit claims.

Five Maxims of Open Ecological Inquiry

Ethics of Artificial Intelligence and Robotics (Stanford Encyclopedia of Philosophy)