Robots versus scientists?
Over the centuries, humans have sometimes had a fear of replacement, whether it is by another human as exemplified in Dostoevsky’s classic short story The Double (or other literature on doubles/doppelgängers) or by machines, most notably robots.
Could we be replaced by robots, artificial intelligence (AI), or cyborgs?
In some ways, robots and in a simpler sense pieces of equipment have taken our places for the better in science doing things more precisely and reproducibly than we can. In science there’s a sense that many things that us scientists do, machines could do better. Imagine in the early days of PCR-like DNA amplification scientists would manually move tubes from one temperature domain to another…
But surely there are somethings that are difficult for non-human scientists to do…
Right? Shouldn’t there be things that are we irreplaceable for doing well? Hmm…in machines versus scientists in the tasks below, who wins out as better?
Cell culture? Apparently machines are available to do this, although expensive.
Western or Northern bots? There are robots for blotting.
Other experimental methods?
Thinking up hypotheses?
Grant or paper writing? I haven’t seen a grant-writing robot or AI yet, but I’m sure it’s coming.
Grant or paper reviewing? I did a post a few years back about robots doing grant review (see its cartoon above).
Mentoring? “Hello, I’m your new graduate (or postdoc) advisor. Please state your name and let me scan your brain to provide the best in computational mentoring…”
Well, that last one in particular seems to require a human touch. Would you want a robotic mentor?
This kind of discussion can sometimes induce anxiety about lost jobs or even Skynet like worries from the Terminator films. In a decade, how much of the lab work will still be human?
If a lot more of it is done by robots or AI or further down the road, by cyborg versions of us, will that be a good or bad thing? With all our human weaknesses and problems manifesting in science with issues with reproducibility, misconduct, and bad mentoring, is a less human scientist and less human science a bad thing? Would science be best served by an army of Spock-like machines that are not even 1/2 human, but not human at all?
Is there such thing as a creative machine?
I did a poll on Twitter about whether machines/AI/robots will take over in the lab and here are the results so far below. If the poll isn’t closed yet, you can go ahead and vote too.
Bing! Working on a piece on possibility of robots & AI replacing many of us scientists some day in the lab. What do you think? How likely?
— Paul Knoepfler (@pknoepfler) October 6, 2017
I recently posted an interview with Dr. Alan Blanchard, who is working to improve cell culture through robotics.
http://biospherix.com/cytocentric/452-cytocentric-visionaries-alan-blanchard-p1.html
Machines are capable of most human physical activities and machine learning algorithms will enable AI to make better decisions than us based on collective global experience. Creativity? Sure, I guess you could set one or two aside with a random number generator to see what they come up with and filter the outcome to see if it better suits the aim of the AI collective.
When AI can do all this, why would they want investigate biological phenomena? More likely they will be more interested in materials science and energy conversion. We should be careful on that “handing over decisions” (aka Skynet) moment!
If we are replaced, probably well replaced we are