Waiting for the cyborg scientists

[This article was first written as an editorial for the October 2019 issue of The Biochemist, magazine of The Biochemical Society. The issue focused on the growing role of Applications of Artificial Intelligence in Molecular Biology].

~~~~~~~~~~~~~~~~~~~~~

From time to time there are innovations whose relevance catches you by surprise. I have to admit that, for me, the applications and implications of artificial intelligence for bioscience is a case in point. I consider that I am usually quite good at “horizon-spotting”, especially in areas with ethical implications, but in the case of AI I got this spectacularly wrong.

Perhaps as a consequence of watching too much science fiction, I languished under the false impression that AI was all about engineers and computer programmers attempting to develop a humanoid being. As such, I thought of AI as being primarily of interest to disciplines more closely aligned to physics. Of course this IS one aspect of work on artificial intelligence. I now know that what I was picturing would be termed “strong” AI (aka “artificial general intelligence” and “full AI”), the attempt to replicate human characteristics such as reasoning and intentionality. (From my recent reading it seems that within the AI field inclusion of consciousness in this list is a moot point, so we’ll leave that there for now).

What I had completely failed to appreciate in my casual dismissal of biological AI was the applications of machine learning to specific tasks, via the capacity to evolve rules. In contrast to strong AI, these are known as “weak AI” or more often “narrow AI”, due to the focused nature of the tasks achieved.