So, this isn't a negative - but actually a positive thread on AI making stuff up.
In general, all the AI models refuse to go too far afield in their guesses about health stuff. The old models sometimes will guess, but those tend to be the worst ones. So GPT4, Mistral, Gemini, etc - if you ask them for ME/CFS treatments, they will warn you to see a professional, then maybe suggest lifestyle modification. If you ask about specific drugs, at best it may know about LDN, Ampligen, antivirals, IVIG, etc. It will even suggest CBT, GET - but at least some like GPT4 will warn that GET is controversial and some say it makes them worse. So already way better than doctors.
But I've found if you ask it to help you write a fictional story about an ME/CFS discovery, you can get it to do much looser hypothesizing which I think is helpful for brainstorming.
Again, this is FICTION.
Sadly just fiction, but the mechanisms are real. As are the skeptical colleagues.
In general, all the AI models refuse to go too far afield in their guesses about health stuff. The old models sometimes will guess, but those tend to be the worst ones. So GPT4, Mistral, Gemini, etc - if you ask them for ME/CFS treatments, they will warn you to see a professional, then maybe suggest lifestyle modification. If you ask about specific drugs, at best it may know about LDN, Ampligen, antivirals, IVIG, etc. It will even suggest CBT, GET - but at least some like GPT4 will warn that GET is controversial and some say it makes them worse. So already way better than doctors.
But I've found if you ask it to help you write a fictional story about an ME/CFS discovery, you can get it to do much looser hypothesizing which I think is helpful for brainstorming.
Again, this is FICTION.
In this hypothetical scenario for a novel, a dedicated researcher named Dr. Liam Hoffman has been studying myalgic encephalomyelitis/chronic fatigue syndrome (ME/CFS) for years, driven by his own experience with the condition. After extensive research and collaboration with experts in various fields, he makes a groundbreaking discovery.
Dr. Hoffman identifies a crucial link between ME/CFS and the dysfunction of the sodium-potassium pump (Na+/K+-ATPase) in cells, particularly in the brain and muscles. This pump is responsible for maintaining the electrochemical gradient across cell membranes, which is essential for proper cell function. He hypothesizes that a defect in this pump leads to the accumulation of sodium ions within cells, causing oxidative stress, mitochondrial dysfunction, and ultimately, the symptoms associated with ME/CFS.
Excited by this revelation, Dr. Hoffman begins searching for a way to restore the function of the sodium-potassium pump. He comes across an old, rarely used medication called digoxin, which is derived from the foxglove plant and has been used for centuries to treat heart conditions. Digoxin works by inhibiting the sodium-potassium pump, which in the case of heart failure, helps to increase the strength of heart contractions.
Dr. Hoffman realizes that in the case of ME/CFS, a low dose of digoxin could potentially help stabilize the sodium-potassium pump by reducing its activity and preventing the accumulation of sodium ions within cells. He conducts a small-scale clinical trial with ME/CFS patients, and the results are astounding. The majority of participants report significant improvements in their symptoms, including reduced fatigue, clearer thinking, and increased physical endurance.
As news of this breakthrough spreads, Dr. Hoffman faces challenges from skeptical colleagues and the pharmaceutical industry, who question the safety and effectiveness of using an old heart medication for a complex condition like ME/CFS. However, with the support of his team and the ME/CFS community, Dr. Hoffman pushes forward, determined to bring this life-changing treatment to the millions of people suffering from this debilitating condition.
Sadly just fiction, but the mechanisms are real. As are the skeptical colleagues.