Dear André,

Thanks for your last comment. I agree that intelligent computers exist that can do a great deal. I am aware that intelligent computers can now undertake many tasks that once required human beings. If we can render these tasks in algorithmic form, intelligent computers can perform them. 

All this is reasonable and well understood. Even those us who do not know how to program such machines are aware that they exist, and we consider the consequences. For example, this is the subject of a recent article by David Brooks in the New York Times

http://www.nytimes.com/2015/09/04/opinion/david-brooks-the-new-romantics-in-the-computer-age.html

Intelligent computers can now perform many routine transactional services in such large and often bureaucratic organisations as insurance companies, banks, or universities, as well as many government departments. These services once required human actors, and organisations grew dramatically in size to meet customer and client needs. Today, intelligent computers that render these services often create higher general satisfaction for customers or clients at dramatically reduced cost to the service organisations.

Intelligent computers also undertake many activities that improve transactional services. Amazon does more than keep track of my purchases and deliveries. Amazon offers interesting suggestions based on my purchasing habits, and it points to interesting items related to the items I buy. 

Many activities and services once performed by human designers are mechanical and transactional. Human designers did not perform these tasks because they required designers to perform them, but because they were necessary tasks linked to the design profession. Drafting in architectural practice is good example of this. I once knew someone educated as an architect who did not want to practice architecture, but did like drafting. He made a good living as a draftsman for many years until intelligent computers replaced him. Vast ranges of jobs and services that once required human intelligence can be performed by intelligent computers, often at lower cost and sometimes at higher quality.

The question I have is whether any computer is able to perform those activities that define design. These activities require attributes that no machine now seems to possess. These are the issues to which Klaus Krippendorff’s comments pointed. The challenges do not merely emerge in non-computable problems and wicked problems, they make equally well emerge in relatively simple problems that require sensitive listening and human interaction. 

Earlier in this thread, Don Norman pointed to situations in which human intelligence and sensitivity assisted by intelligent computers can perform many services better than either alone — “augmented intelligence.” There is evidence for this in medical practice where physicians can perform better diagnostics with computerised assistance than they can perform without such assistance. In legal practice, we also know that computers can sort through documents and locate relevant precedents more rapidly than a human lawyer. But medicine and law require some skills that computers do not yet show.

In asking whether computers — or machines of any kind — can design, I am asking whether any computers possess the capacity of intention or making a choice to do something without being programmed to do so. I ask whether a machine can have purpose or form purposeful or teleological goal-orientation as contrasted with doing so when a human operator or partner sets the goals or establishes the purpose. And I ask whether a computer or intelligent machine can plan independently, as contrasted with executing a plan at a high level of skill.

There are other capacities involving response, judgement, and emotion that computers do not yet demonstrate.

I am aware that many human beings lack these qualities or possess them at lower levels than other human beings. That is why we see greater and lesser practitioners in every service profession, and that is why some people have no aptitude at all for some professions. 

We are also aware that some human beings function in a mechanical way, and we know that some human beings seem to have little sense and little capacity for reason. 

The question is not whether a computer can emulate or replace these human beings. It is whether a computer can replace the functions and skills we require in an excellent designer, an excellent physician, or an excellent lawyer. It is the judgmental and responsive nature of professional service that led Herbert Simon to speak of “design sciences” that include such professions as medicine, law, and management within his concept of design. In addition, there are questions of ethical engagement and contextual heuristics in which professionals reach out through active inquiry         in ways that are not yet accessible to programs or algorithms.

These are the issues that some of us have been raising. I understand and agree with you on the power of machines with respect to cognitive functions and many tasks that once required human beings. In relatively reasonable awareness of what intelligent computers can do, I am asking whether they can yet perform at the level of a serious service professional. It is in this sense that I have been asking, “can machines design?” without agreeing that they can design in the full range of intention and engagement that we expect in serious design. It is clear that human beings can now program a wide range of subsidiary design skills, and it is clear that machines can now perform these subsidiary skills, often better than human beings and nearly always at lower cost. 

The question is whether any machine now possesses the core skills that would enable a machine to perform the *full* range of activities we expect of a professional designer operating at a high level of expertise. This designer may well be assisted by junior designers or possibly by machines, but I am asking about the full range of expert design skills and not the lesser skills that may be required to deal with any specific job or challenge.     

Yours,

Ken

Ken Friedman, PhD, DSc (hc), FDRS | Editor-in-Chief | 设计 She Ji. The Journal of Design, Economics, and Innovation | Published by Tongji University in Cooperation with Elsevier | URL: http://www.journals.elsevier.com/she-ji-the-journal-of-design-economics-and-innovation/

Chair Professor of Design Innovation Studies | College of Design and Innovation | Tongji University | Shanghai, China ||| University Distinguished Professor | Centre for Design Innovation | Swinburne University of Technology | Melbourne, Australia

--

André Neves wrote:

—snip—

the original question "can machines design?" clearly brings the need to define both terms, "design" and "machines".

In this debate, we spoke a lot about the term "design", since we are all "experts" in design. In general, the definitions of the term are mature in this group and lead to some consensus.

However, we talked very little about the term "machine". Some attempts to define the term "machine" are superficial, more intuitive and less academic.

In Computer Science we differentiate traditional computers and intelligent computers. Traditional computers are dedicated to tasks that machines can do easily and humans has difficulty. For example, perform complex calculations or compare large amount of data.

On the other hand, intelligent computers are dedicated to tasks that humans do easily and the machines has very difficult to perform them. For example, choosing a gift for a particular person or selecting the right clothes for a meeting. Or maybe, "design".

We still do not have a lot of examples of machines as well, but the Computer Science has advanced significantly in the last 50 years in research with AI.

—snip—


-----------------------------------------------------------------
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
-----------------------------------------------------------------