Researchers in London are claiming to have created an actual robot scientist. Much like a human scientist, the robot is capable of reasoning and carrying out experiments on its own. Ross King for Aberystwyth University has christened this robo-scientist, Adam. So far, Adam has performed experiments on the metabolism of yeast and has since become the first robot to make a scientific discovery. Adam was able to detect the gene in yeast that is responsible for the production of an amino acid called lysine. This amino acid is especially important because it is vital to growth. How did it do this? “The robot, called Adam, was able to work out where an important gene would be located and to develop experiments to prove its theory” (Smith). This discovery and those expected to be made in the future, are likely to be important to creating new treatments to illnesses. Adam’s discovery in particular will be useful when treating fungal diseases such as the common Athlete’s foot. This is because the gene responsible for growth can now be identified and disabled. Though robot scientists will without a doubt become important, it is unlikely that they will replace human scientists. Instead, robo-scientists will work alongside humans. They will be used to, “…carry out large numbers of repetitive tests that in a person would induce boredom and loss of concentration” (Smith) and to record very specific details. A new robo-scientist by the name of Eve is expected to be switched on soon. It is only a matter of time before these robot scientists are no longer a rarity.
For more information:
http://www.reuters.com/article/scienceNews/idUSTRE53167K20090403
Smith, Lewis. "Robot scientist ." Times Online. 03 Apr 2009. The Times. 5 Apr 2009 http://www.timesonline.co.uk/tol/news/uk/science/article6024880.ece
I could use a robot for some of my own research, actually...
ReplyDeleteIndeed, it occurs to me that in many fields, scientists already have helping "robots" -- but the robots aren't physical; they're automated programs that work side-by-side with the scholar/scientist, doing the sorts of things that humans would likely screw up out of distraction and boredom. That's the sort of thing I need help with, and so it's not surprising that lab scientists would want one, too.
I wonder how long it will be before we hit a kind of "singularity" and the AI-type systems start churning out more discoveries than we can easily keep up with. The world five years from now could, potentially, be completely alien to the one we have now, if we ever hit an exponential progress curve. I've heard some speculation that this could happen, but don't know how seriously to take that. What do you think?
I think it's unlikely that we'll have AI scientists who can perform a wide range of tasks for a while yet. The main stumbling block as I see it is that we have to be able to program either a) an effective routine that let the robot learn or b) everything a scientist might do in a particular instance. And how do you account for inspiration? So, for now, a) is the only option.
ReplyDeleteI think the word 'theory' is slightly misleading. It implies that robot is actually thinking in a human like manner, when in fact is is deducting (something computers do very well) based on available facts within a set framework.
Still, interesting article.