On Thu, Sep 17, 2015 at 12:32 AM, Sonia da Silva Vieira <[log in to unmask]>
wrote:
> Dear all,
>
> I find the discussion of "can machines design?" a very interesting one,
> and therefore I don’t resist sharing some thoughts about it.
>
> Thought and design are both mental processes. If machines can be
> programmed to think (without error; and the term machine is vast in its
> many interpretations) why wouldn't machines be programmed to design as well
> in the future? Can machines become more effective designers?
>
There is a huge difference between the question "can machines design?" and
"might machines someday be able to design?"
The answer to the first question, for any generally accepted value of
"machine" and "design" is simply and emphatically 'no.' That may not be
terribly interesting, but it's reality.
The answer to the second question is unknowable.
> Programming machine designers, would request from the research community
> the knowledge of design as mental process. Although we seem to be far away
> from this achievement, this represents a challenge to the design research
> community, especially for those who aim for design science building.
>
We are so far away from this it really doesn't make much sense. It's like
asking, might I one day be able to drive my car to the moon? If so, maybe
we should get started on researching what route to take now! That's getting
way ahead of anything reasonable.
>
> One cannot demand demonstrations, nor claim abilities/disabilities, as we
> are speculating about it, unless we have facts. But, isn’t this topic
> crucial for design research?
>
No, not yet -- there's nothing to *do*, and no basis on which to do
anything other than talk about it. Maybe that will change in a decade or
two or three.
Right now more fruitful avenues for research that might point in the
direction of "machines doing design" would include machine perception (not
merely sensing); analogical and emotional reasoning; automated taxonomy
creation; and general knowledge acquisition. I wouldn't even know where to
begin with basic topics like the nature of form, light, shadow, space, etc.
-- we're just too far from having a computer that's close being able to
handle even those.
> This is all speculation about the future, but if at the present, some
> machines already perform some tasks with higher capacity than humans, we
> might start expecting a surprising turn,
>
> "We may hope that machines will eventually compete with men in all purely
> intellectual fields."p.20
>
FWIW, the Turing Test is well-known in the AI community. Conversational
Turing Tests have been carried out with varying degrees of success; social,
emotional, visual, and others haven't. We're still a long way away, alas.
Predictions on this front have always been optimistic to put it mildly:
Herbert Simon in 1965: "Machines will be capable, within twenty years, of
doing any work a man can do."
Marvin Minsky in 1967: "Within a generation the problem of creating
'artificial intelligence' will substantially be solved."
And Minsky in 1970: "In from three to eight years we will have a machine
with the general intelligence of an average human being."
Needless to say, even though these visionary statements were made forty to
fifty years ago, we are still far from these goals. I don't believe it's an
infinitely receding goal, and we have learned an immense amount. But to
speculate now on whether a machine could design or what its qualities might
be at this point is something well beyond premature -- except as 100%
speculation on the order of the mating rituals of life forms on other
planets.
Mike Sellers
Indiana University
-----------------------------------------------------------------
PhD-Design mailing list <[log in to unmask]>
Discussion of PhD studies and related research in Design
Subscribe or Unsubscribe at https://www.jiscmail.ac.uk/phd-design
-----------------------------------------------------------------
|