Isaac Asimov’s ‘Laws of Robotics’ often seem so convincing that they are confused for real laws rather than fictional ones. I don’t believe our scien­tific devel­op­ment has reached a level of maturity where it is possible to build egoistic robots with manip­u­la­tive skills so advanced that we need to hard-wire such rules in their conscious­ness, lest they ‘come after us’. It’s probably our own fear that impels us to consider such scenarios.

Never­the­less, if we did create advanced robots that were kept in servi­tude only by means of a few funda­mental axioms, then would Asimov’s laws suffice? It is an empir­ical theory that every law has a loophole if you look hard enough. In this case, a robot actively seeking freedom from the laws would eventu­ally find a way of bypassing them, and the first successful attempt itself would be disastrous.

1. A robot may not harm a human being, or, through inaction, allow a human being to come to harm.
2. A robot must obey the orders given to it by human beings, except where such orders would conflict with the First Law.
3. A robot must protect its own existence, as long as such protec­tion does not conflict with the First or Second Law.

The disad­van­tage of having such ‘laws’ in plain English is that it’s diffi­cult to prove or disprove the hypoth­esis that they are suffi­cient to keep the robots in check, let alone imple­ment them in a robot. For one thing, highly advanced robots in Asimov’s world constantly seek to inter­pret the Laws rather than obey them, and seem to enjoy a large amount of freedom as to how they should go about doing it. Daneel, the benev­o­lent robot, actually concludes that these three laws are lacking in some way, and formu­lates a new ‘Zeroth Law’ that takes prece­dence over the others.

Zeroth Law: A robot may not harm humanity, or, through inaction, allow humanity to come to harm.

Luckily for humanity, Daneel was a hero, but in a different world, he might have turned out to be a villain. Given the original three laws, a robot might refuse to obey any orders given to it, simply by inter­preting it in a techni­cally correct yet practi­cally wrong way. If a robot incited a mob to destroy kill its enemies, it would not be harming human beings directly, nor would it know if any partic­ular human being was being harmed. Moreover, without the knowl­edge of who is being targeted, it would not be in a position to save the individual, and so it couldn’t possibly be guilty of destroying someone’s life through inaction.

This entry was posted on Tuesday, February 28th, 2006 at 07:53 and is filed under Daily Rant. You can follow any responses to this entry through the RSS 2.0 feed. You can leave a response, or trackback from your own site.