Saturday, June 06, 2009

A Robot Code of Ethics?

May 10, 2007

Dear Glenn McGee,

I read your recent column "A Robot Code of Ethics" (The Scientist, May 2007) with interest, as I have reflected on the same issues in my paper "A Jurisprudence of Artilects: Blueprint for a Synthetic Citizen" (November 2001).

First let me say it pains me whenever anyone mentions Isaac Asimov's "Three Laws of Robots" as if they are to be taken seriously. They were devised by a science fiction writer to sell books, and I do not believe they provide a starting point for any serious legal analysis. For example, if a highly valuable, deeply sentient robot were under attack by a low class human criminal seeking to strip and sell its memory to buy drugs, I would read the First Law as saying the robot has no right of self defense and must let all the wealth of its experience and relationships be lost.

Already we are getting nowhere. These "Laws" may reassure a nervous, uneducated human public of perpetual human supremacy, and perpetual robotic servitude, but they alas don't enlighten anyone about the nature of consciousness and social relations. Hence any mention of Asimov's Laws should be only to dismiss them as useless.

Turning to the question of what could work, I am unimpressed by your proposal for a code of ethics, since (as I discuss in my paper) we already have a legal system to govern such issues. If a robot is (a) below some level of sentience, it is a machine and if it commits torts or breaches contracts its owner will be responsible, and (b) above some level of sentience it might achieve emancipation as an independent legal persona, such as by becoming incorporated, as I propose in my paper.

There is certainly a need to discuss these issues, and I commend you for doing so, but I believe you've got it wrong. Your piece seems to assume that robots will be outside the current legal framework, and a new code of ethics needs to be devised, whereas in reality there is no new zone of conduct needing to be regulated. Any code of ethics will either (a) protect the owner from being sued, or (b) permit the new citizen to function more effectively and/or obtain liability insurance for itself, if required by law or business practice, to get people to become comfortable in dealing with it.

Certainly more can be done to define these issues, but it would be desirable to get people with standard legal training involved, to get everyone to the initial realization that existing laws already address most, if not all, of the major points of concern.

Frank W Sudia, JD
San Francisco, CA

Labels: , , , ,

Getting Robert Rosen Back in Print

June 2, 2009

James Rutt, Trustee
Santa Fe Institute
1399 Hyde Park Road
Santa Fe, NM 87501

Dear Jim,

Last I saw, you are Vice Chairman of the Board of Trustees at the Santa Fe Institute, making you a leader in the field of complexity studies. With leadership comes its burdens :-), so I have a request.

Robert Rosen, the mathematical biologist from Columbia University, was (with his teacher) a founder of relational biology and pioneered the application of Category Theory to biology. He gave at least one presentation at SFI, and got along with your guys. But he died prematurely (diabetes), after being vilified by traditional mechanistic biologists, leaving his daughter Judith Rosen as his literary heir.

I tried (without success) to reach Judith about getting her late father's critical works back in print, especially his landmark Anticipatory Systems, but she apparently has become depressed about the whole thing, and openly wishes her father's work would generate revenue for her, which I doubt is in the cards.

Pioneering scientists often do not live to see their work widely accepted. Gregor Mendel wasn't famous until decades after his death, nor was Alfred Wegener who invented plate tectonics. It goes with the territory.

Rosen was a genius contributor to the mathematical theory of complexity, but many of his works are out of print (Anticipatory Systems is not available from any book dealer at any price), and others (such as Life Itself) are in print, but with ridiculous numbers of errors and no errata sheet. Judith says (plaintively) that Columbia University Press was informed and given a corrected copy, but when the paperback edition was published all the errors (and there are dozens) were still there.

To cross the great divide and arrive on the other side of scientific acceptance, it would be most helpful if younger scientists could have ready access to his works. This could cut X decades off his hibernation time until he is rediscovered and deified.

Amazon's non-entry for Anticipatory Systems has a button that says something like, "if you own the rights and want to see this back in print, click here." Likewise Google Books has become a huge force in publishing, scanning massive numbers of works. There are plenty of ways this problem could get solved.

I would like to own Rosen's key works, with errata sheets as needed, which is basically impossible at present, unless maybe you can use the imprimatur of the SFI to persuade Judith to authorize the their reissuance, as well as make available such errata sheets as she currently has. I'm sure either service will pay her a few bucks of royalties. (I would gladly pay her 100% cash for a PDF copy.)

Given that complexity has been a slow burn, it could be a more general (and low cost) SFI goal to make a list of other authors whose works should be "kept readily available" for rediscovery by so-called mainstream science. I nominate Rosen as the poster child. (Anticipatory Systems is like 450 pages long, and can be read at the Stanford University Library, but who wants to stand there and xerox it? Not to mention copyright issues.)

Dunno where this would fit on your radar screen. As someone once said, "you make your predecessors famous," which I hope to help do for Rosen, but I'm not there yet, and the nonavailability of key his works is not making it any easier.


Labels: , , , , , , ,