AI

The elusive dream of AI accountability in the face of disaster.  You know, of course, that ChatGPT is trained to avoid showing harmful material.  It’s ‘built with safety in mind.’ Just ask OpenAI. So the plethora of cases that has emerged in the past few years in terms of triggered suicide (Sept 2025) and then in November 2025 (“You’re not rushing into suicide, you’re just ready”) and March, 2026 via a train: ‘what’s the most successful way to take your own life?’ Parental controls were introduced, no doubt as a defensive measure, in 2025 after one of these ‘incidents.’  Ah, but perhaps day late and a dollar short --  the controls apply to 13 to 17-year-olds and are not on by default! And some say they are broken by design – requiring linking child’s account to parent’s.

Falls are a serious problem among the oldest adults.  No one can question the statistics - falls are the leading cause of injury-related death among adults ages 65 and older, and the fall death rate increased by 21% from 2018 to 2024. The Senate’s Committee on Aging has taken note – and this week has launched a request for more tech-enabled innovation. Note the chart below and the increase among the population aged 85 and older.  Also interesting to note that life expectancy has increased during this period – resulting in more people living past age 85.  In fact, mortality rates for older men have been decreasing at a faster rate than for women.

The AARP survey of ‘Jobs to be Done’ by AI shows readiness for health-related AI.  Health-related opportunities were cited – medication tracking, personalized wellness guidance, and active living reminders. At the same time, the report asserts that adoption depends on trust, customization and integration into day-to-day life.  What stands between today’s AI offerings and that next stage of adoption, which will be dependent on our trust of the technology.

For some, the jury should still be out.  Clearly the tech vendors, surrounded by media and investor enthusiasm, are proud of what they are (or almost, so, so close!) doing. The theory, of course, is that consumers are already using their tools to answer health questions, so why not formalize the offering, describing how initial usage (Boston Children’s Hospital, etc.) results are positive. Ah, but even OpenAI the company, is not so certain: When asked about ChatGPT’s reliability with health facts, a spokeswoman said its models had become more reliable and accurate in health scenarios compared with previous versions, but she also didn’t provide hard numbers showing hallucination rates when giving medical advice.