Having kept up, more or less, with goings on in AI, it seems quite clear to
that there's no way to remove the human from "designing" for now. I don't
doubt that someday there will be artificial consciousness, but I don't think
we have to worry about it happening anytime soon.
Still, I note that there have been some interesting sidebars over the
years. EG: I recall a case-based reasoning system (I think it was CADET)
that was able to "design" (perhaps "re-invent" is a better word) a proper
faucet with both hot & cold water taps based only on the rules of heat
transfer and of how seesaws worked (that is, it "knew" nothing about faucets
and fluids). I would be hard-pressed to discount this case as one of a very
primitive design done entirely by a machine.
I prefer to think of "agents" that are complexes containing a human and a
variety of other things, including computer tools, pencils, etc. As things
seem to happen today, removing the human element from the agent wrecks it
entirely, but removing one or more of the other things can substantively
change the nature both of the design activities and of the design. There
are things that have been designed that could have never been designed if it
were left solely to the unaided human designer, no matter how many of them
you threw at the problem.
So it's not a matter of whether design is or is not a human activity, but
rather the degree to which it is so, and the size of the domain of design
problems that will submit to a designed solution.
2 cents on a muggy August Saturday.
2009/8/15 Ken Friedman <[log in to unmask]>
> At some point, I may reply at length to Terry's post and the rejoinders. A
> short note must do for now. There is an epistemological and ontological
> confusion in Terry's notes, and I think that both Robin Hodge and Gunnar
> Swanson have identified some of the problems.
> The short argument is this -- people design, machines do not. Design is a
> thought process. If design is a thought process, machines can't substitute
> for the design act. Rather they provide useful tools, and some aspects of
> research have made the tools work better. Some machines take over process
> acts and replace skills. But as Gunnar notes, using a computer to design a
> page is quite different to designing the type we set. Of course, computers
> also assist those who design type, but the point is clear. One is a
> task that junior designers and even amateurs can now manage. The other
> remains an expert-level skill, even with a computer and the proper
> Nigel Cross's work on designerly ways of knowing examines what it is to
> design. Beyond this, it is worth noting that designers in many fields use
> design thinking to design many kinds of artifacts and processes.
> Design research involves many projects and problems -- designing and design
> acts are merely one range. Materials, reception, meaning, emotion may all
> subjects of design research. In the sense that a science is an organized
> body of knowledge, the research we do in design schools means that design
> schools are becoming scientific -- but science does not mean that computers
> replace designers any more than computers replace researchers. Computers
> assist and speed up the work that researchers do in mathematics and
> too, but they do not replace mathematicians or physicists.
> For the rest, I shall wait.
> But please -- I request again that those who post trim the prior posts when
> replying. I've read Terry's full post three or four times now. Those who
> daily digest will have a sorrowful time with this thread.
> Ken Friedman, PhD, DSc (hc), FDRS
> Swinburne Design
> Swinburne University of Technology
> Melbourne, Australia
Filippo A. Salustri, Ph.D., P.Eng.
Mechanical and Industrial Engineering
350 Victoria St, Toronto, ON
M5B 2K3, Canada
Tel: 416/979-5000 ext 7749
Email: [log in to unmask]