Can AI Find a Cure for Alzheimer’s?

The big, big problem. No cure so far for Alzheimer’s, none in sight. We need to try something really new. How about ChatGPT?

Physician Cure Thyself

I’m sure I don’t have to tell you that AI, and in particular the product-du-jour is ChatGPT from OpenAI. It has already achieved some outstanding feats, such as being able to write fluent, convincing and constructive essays and papers good enough at least for an SAT paper if that’s your gig. It can create original images when requested of anything at all, even down to the genre and style you wish. It can even do this for videos. And it can hold a chat with you that might enthrall you or scare you out of your wits.

Naturally, the Luddites are already up in arms about the implied diminution of human capabilities. But right now my question is more short-term and not at all philosophical, to wit: can it find a cure for Alzheimer’s?

I think Alzheimer’s is the perfect test for AI. That is, humans haven’t been able to find any form of treatment, even palliatives let alone a cure. Despite spending billions of dollars we are no closer to a cure. Recent Alzheimer’s drugs have failed in clinical trials There’s no convincing research in progress that looks like to effect either a cure, or a palliative. Our only prescription is to recommend lifestyle changes. What would they be if you have it already or well on the ways towards it?

Alzheimer’s is a great paradigm for the effectiveness of AI because its unsolved and probably unsolvable on present trajectories. That’s because despite the involvement of the smartest scientists on the planet we still are seemingly no nearer to getting anywhere. Solve the Alzheimer’s problem and it clear that there’s a new horse in the race to go beyond the limited capabilities of the human brain.

Part of the problem might well be the explosion of neuroscience data and the bewildering assortment of new techniques in biochemistry, These days there’s stuff like CRISP-R (in various flavors these days), genetic databases, monoclonal antibodies and so much more. It could be that humans just can’t cope with this volume and its hyper-complexity. How do you deal with that?

A Manhattan Project for Us All

Here is my proposal. We need to launch one or more projects to find a treatment for Alzheimer’s using AI. The NIH and other research-funding agencies should make this a top priority and take some money away from approaches that haven’t been productive so far such as pursuing amyloid plaque approaches. The aim would be to get AI to target completely new approaches that haven’t been thought up or tried so far.

Do We Need the Know-Nothings?

I think that the fund granting individuals and agencies need to adopt some new approaches that are counter-intuitive so that we know we aren’t just duplicating old approaches using different labels. These might include:

• No background in neural science: the project personnel don’t need any background in neural science, in fact it might even be regarded as a liability. Let’s avoid approaches from professionals who haven’t solved the problem so far.
• Not professional researchers; project personnel don’t necessarily need to be professional researchers in any disciple e; let’s get some talented amateurs involved who don’t have an ax to grind(such as getting tenure, more publications of a Nobel prize).
• Don’t need Ph.D.; for much the same reason project peoples should have to possess an eligibility qualification such as a Ph.D. or similar.
• Come from totally different disciplines; we want some people with different backgrounds in apparently unrelated disciplines. How about MacArthur Fellows in medicine, or Fulbright scholars in cultural anthropology?
• Already have some AD; could it be that some sufferers of Alzheimer’s have insights that you wouldn’t have if you haven’t suffered it already? Particularly ones that might have some medical knowledge (totally opposed to my recommendations above e I know, but worth trying anything, right?).
• Support repurposing of already approved drugs: There’s loads of already approved drugs out there but the drug companies have powerful financial reasons not to do any research into existing approved drugs. How about using AI’s powerful capabilities to review the thousands of drugs out there to assess their likely impact on AD? Lets see if we can more value of the expenses we have already ploughed into medical research.

The Alzheimer’s scourge is an existential tragedy for millions of people in the US and worldwide. Let’s get serious about doing something about it!