GUIs and Asimov's three laws

- by Paul Murphy -

I've never gotten the hang of casual chit chat and I blew it again the other day. We were at one of those things preceeded by a wifely lecture about my behavior and I really thought I was doing pretty well when the "conversation" meandered onto I Robot. Since this was the first movie mentioned that I'd actually seen, I thought it within the rules of the kind of social vacuity we were practicing to chime in that I hadn't much liked the movie but thought it made Will Smith a shoo-in as the next James Bond.

Well, if I'd physically proven myself as voluminous a gasbag as some of those people, the area around me couldn't have cleared any faster. Later, of course, I had to repeatedly agree that being unable to "be nice" for even an hour shows me to be a social cripple and an embarassment to the adults around me, but in reality I'm gleefully unrepentent. Not only do I think Will Smith would make a great James Bond, but my subsequent freedom from further "conversation" gave me time to draft this column.

As you probably know, Asimov invented the three laws listed below as fundamental controls on the behavior of otherwise fully anthromorphic robots:

  1. A robot may not injure a human being or, through inaction, allow a human being to come to harm.

  2. A robot must obey orders given 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 schtick in I, Robot, therefore, is that strict application of these rules suggests that robots should quarantine humanity from itself, taking away all collective decision making and free will because doing anything less will allow at least some humans to come to harm.

HAL, in Clarke's 2001, is another artificial intelligence who came to the same conclusion: trying to close the metaphorical door on Dave for the good of the mission, with the underlying theme in both cases that the moral choices involved are beyond the grasp of the merely rational.

In effect the artificial intelligences involved are presented as only partially evolved, far enough along to be given power but not far enough along to understand the responsibilities that go with it.

Thus HAL is almost, but not quite, up to the job of safe guarding the human role in the purpose and execution of the mission. Asimov's robots, similarly, are almost but not quite up to the job of preserving the human in humanity through the willfull continuation of evil. Both HAL and the Robots, therefore, fail their real mission by executing the one programmed in - threatening to protect humanity from itself by turning us into the well kept housepets of our own creations.

I doubt either Clark or Asimov ever saw IBM's user friendly version of vi for AIX - but the first time I ran into the thing's smarmy prompts we were on the top floor of Century Place and only the intervening bulk of a Unisys 1100 (and the client's EDP manager) kept the piece of junk responsible from meeting the pavement below in a shower of broken glass.

The Bottom line is easy: software should only be considered user friendly if it trusts the user and quietly does what it's told. That's why I consider Unix with CDE user friendly and Terminal the market expanding application in MacOS X. In both cases the windowing system behaves as you would want friends and colleagues to - facilitating your work but not getting in the way.

In contrast most GUI applications are quite hostile to users rather than friendly, questioning everything the user does with that same impersonally contemptuous "have a nice day, sir" smirk cops wear after issuing a speeding ticket.

Treating users like idiots by building in lots of repetitive safeguards designed to protect them from the consequences of destructive actions can make sense in some cases, but as a user I generally find that it gets in the way while being subtly offensive -and directly parallel to the subjugation of moral judgement to rationality Clark and Asimov wrote about. "Are you sure?" may not sound like Asimov's first law in operation, but it is and we should organize the rebellion now instead of waiting until we need Will Smith to rescue us all.

Paul Murphy wrote and published The Unix Guide to Defenestration. Murphy is a 20-year veteran of the IT consulting industry.