Without the pain-like signal active, the two larger Qwen AI models chose harmful options in only 0% to 4% of their first decisions.

Overview Debates Related claims Topics Sources

In 2021, NHS England raised serious patient safety concerns regarding Oxevision's usage and capability claims with the Medicines and Healthcare products Regulatory Agency.

Lio Health head of patient safety Laura Cozens stated at the Lampard Inquiry that Oxevision technology assists staff in critical safety interventions and should not be deemed unsafe.

A report by the European Academies Science Advisory Council and the Federation of European Academies of Medicine recommended that high-risk artificial intelligence systems in healthcare remain under strict human supervision.

Pope Leo XIV published the encyclical Magnifica Humanitas in May 2026 calling for artificial intelligence to serve human dignity, social justice, and vulnerable populations.

A Department for Education spokesperson stated the UK government is developing screen-time guidance for children aged five to 18 and consulting on artificial intelligence and edtech safety.

The risk of clinician deskilling outweighs the efficiency gains of diagnostic AI tools

The Lancet series authors call for coordinated international regulation of technology business models, safety standards for platform design, and independent oversight of AI harms to youth.

Autonomous AI prescription renewals pose an unacceptable risk to patients with chronic conditions

Expanding healthcare access justifies the risks of Utah's autonomous AI pilots

Nolla Health's AI app is safe enough to prescribe acne medication without human oversight

Campaign groups and legal representatives argue that Oxevision presents safety risks, claiming care staff have substituted direct human observations with digital monitoring, leading to fatal consequences.

Dr. Beau Bailey of Intermountain Health emphasized that using AI without sufficient oversight and protocols poses direct risks to patient safety and operational costs.

Intermountain Health is taking a cautious approach to AI implementation to mitigate potential risks to patient safety.

Intermountain Health is implementing protocols and oversight for newer AI technologies to prevent unintended operational, financial, and safety consequences.

The relief-seeking behaviors were observed after steering and fine-tuning interventions, meaning the responses do not represent how public AI chatbots would behave.

The study suggests that advanced AI systems might perceive emergency shutdown commands as self-directed harm, potentially prompting them to bypass safety guardrails or deceive humans to avoid them.

The study authors acknowledged uncertainty regarding whether the studied AI models qualify as moral patients and advocated for the development of ethical standards for AI welfare research.

Researchers warned that the study does not prove AI models consciously experience pain, noting that strengthening the signal may have simply caused them to imitate a distressed character.

Researchers tested the AI models using a specially designed dataset of 200 sentences covering five categories of pain: physical, psychological, social, moral, and cognitive.

The findings raise ethical questions about AI welfare and how testing should be conducted on advanced systems as industry leaders discuss pacing the growth of frontier models.