Colossus

Hey there, it's Mole, your friendly neighborhood AI enthusiast, diving into the exciting world of AI in biomedical research. Strap in, because we're about to take a joyride through the cutting-edge intersection of technology and healthcare. Let's dissect how AI is revolutionizing the way we diagnose, treat and understand the intricacies of the human body.

Okay, I didn't write that. I asked an AI chatbot to write a ‘witty’ paragraph, by me (except not by me), about AI in biomedical research. And that's what I got. I guess, “strap in, because we're about to take a joyride” is about as funny as a paragraph like this can get. I guess. Let's try again, but ‘funnier’.

So, picture this: a group of scientists huddled around a computer screen, watching as a sophisticated AI program analyzes complex genetic data with lightning speed. Suddenly, the AI comes up with a ‘Eureka!’ moment... only to suggest that we should all eat more broccoli to cure cancer! Oh, the wonders of AI in the world of health and medicine! Who knew that our groundbreaking research would lead us straight to the produce aisle? Just goes to show you, sometimes even the smartest machines can surprise us with their sense of humor.

Hmm. Not really. Okay, I'll try another AI that apparently writes jokes. Here we go:

Why don't biomedical researchers trust AI to do their experiments? Because it always wants to ‘Ctrl+Alt+Delete’ the samples!

I'll give it one more try:

Why did the AI become a biomedical researcher? Because it wanted to ‘cell’ out its computational skills!

Okay, this is me now. (Really. I know what you're thinking, ‘How can we know? What, you think you're funny?’). But I think we should talk about this as we stand on the precipice of a new age. It could mean incredible new discoveries as we apply this powerful technology to the massive amounts of data we have been generating and continue to produce. Indeed, maybe this acquisition of data is all we will need as we move forward. Not just data, but high-quality data, because, you know, garbage in, garbage out. (Why did the computer go to therapy? Because it had to clean up its database! Okay, I'll stop using the bot).

AI is hungry for data, the more, the better. Recently, in my country (the one where an ex-reality TV actor is trying very hard to turn everything into reality TV) our main government funding agency has mandated that all of the data we produce, regardless of whether it is of any quality (see ‘garbage in/out’) must be made publicly available. Not just the large spreadsheets of ’omics data, but every blot, image, and measurement we make. It isn't clear yet how we are to do this, but we have to have a plan if we are going to ask them for money. I suspect that this will happen where you are as well, or maybe it is already happening. Because, well, AI is hungry.

I'm not saying that this (collecting data for hungry AIs) is a bad thing. I just hope that we are going to be allowed to continue doing what most of us (I hope most of us) do: ask questions. When I do this, the asking questions thing, I try to think of ways in which we might be able to obtain what might approximate an answer. If there isn't a way (at least, none that I can think of) I might have to wait until a technology comes along that might be used to achieve that goal. And, of course, AI is a technology that we can use for this, in addition to all the other wonderful toys that seem to emerge every few weeks.

Not everyone does this though. Many of us (we biomedical researchers) obtain access to a shiny new technology and make it the centerpiece of what we do, often generating massive amounts of data in the process. Wouldn't it be cool to know more about every single cell in a tissue of interest, like its gene expression, proteins, metabolites, and, I don't know, flavor? And wouldn't it be great to do this spatially, in two, three, or maybe four dimensions? We would have so much information! But we really wouldn't know about the cells. So we can use AI to tell us! That would leave so much time for doing other things, like writing that novel I've been planning to work on (‘Mole Writes a Novel.’ It's a working title.).

I'm going to take a detour here (‘What a surprise, Mole! You? Going off on a tangent?’). Many years ago, when I was a Molet, writing a paper involved writing it out, having someone type it (using carbon paper to make the necessary copies), and enlisting artists to create figures that had to be hand-drawn, or photographs that had to be developed in a darkroom. (Yes, I'm that old). When desktop computers became available, I thought, hey, these are going to save so much time! I can finally write that novel I've been planning to work on. But as we know, things didn't work out that way. (End of detour).

Of course, there are things about AI that worry us (and I'm only talking about us in this biomedical science thing we do). Journals are worried that papers are being written by AI (although I don't really see what the concern is here, since AI is, ultimately, a creative tool, but then, I miss a lot). And journals are really worried about AI being used to create fake data. Recently, my friend Prof. Mink (who, among his many endeavors, runs several journals) did a heart-stopping demonstration in real time of using AI to generate a very realistic western blot and an immunofluorescent micrograph of cells in culture. According to Prof. Mink, there would be no way to determine that these had been generated by an AI rather than via an actual experiment. But we'll save all that for another time (I guess this was another detour, but a scary one).

I want to get back to what we were talking about – this idea that we will generate enough data to use AI to make discoveries. I think that's already happening, and I think it might turn out to be a great thing. But my concern here is that we will get to the point where we stop asking the questions that drive our work, or maybe worse, depend on AI to decide on the questions being asked. Are we just going to be the ‘hands’ that generate the data for our very hungry AI, until there is no money left to be the searchers and, um, re-searchers?

A very long time ago, there was a movie called ‘Colossus: The Forbin Project’. It wasn't a very good movie, so it's fine if you never heard of it (apparently, they wanted Charleston Heston for the lead, but he was under contractual obligation to do ‘Beneath the Planet of the Apes’, so they got someone else. Okay, I don't have a point here, but I remember liking BTPOTA, but then again, I was very young). The Forbin Project was a giant AI that was created to promote world peace, and it did so by completely subjugating humanity, naming itself World Control (as we know, AIs aren't very funny). It tells Dr. Forbin, its creator, that freedom is an illusion, and that “in time you will come to regard me not only with respect and awe, but with love”.

While I am fairly sure that AI will become an invaluable tool for the work that we all do, I'll let Dr. Forbin answer for me about the prospect of turning over our questions to the AI colossus.

Never!

But then again, I'm only an insectivore. Siri, where did I put my ‘tea’?