(…continues from the previous entry)

Ubois:
So any other final thoughts on how to promote the concept of responsibility and innovation, you know, in public opinion and with practitioners? I’m also looking for pointers of other people to talk to institutions that are or might be pursuing this line of inquiry.

Twidale:
I think that there’s a lot of discussion now on sort of the environmental consequences of what’s being done. This is huge in Europe, but I don’t want to dismiss it in the US. I think it’s much more visible on the coast, but, you know, it even reached Illinois, so it’s – we’re normally a bit behind the times on these kinds of things. So I think that, considering the environmental consequences of what it is that you’re doing, and that requires as much as anything, a different mindset that can be extended to other things. So, if you’re in a purely business world, and if you’re working for a big corporation, it’s about making money. It’s a very clear measure, an objective measure, is this good or bad, how much money is it going to make. And the problem that we have to a system is that it does not incorporate dollar value on all the things that matter to us. Carbon emissions being a classic case, pollution, and all these other things.

Ubois:
Let me externalize my cost.

Twidale:
Absolutely. So you have all that issue and a lot of the debate in environmentalism is can you bring some of these costs back into the market place, call them taxes or whatever. But once you’ve set that precedent, you can do it about a whole load of other things too. And so, it becomes reasonable to say, okay, what are not only sort of the short – even in pure economics, there’s short term gains and long term costs. And a good company should be thinking about both. And a company that’s purely led by its stock price often only thinks about the short term gains, does not invest in R&D, for example. So, even with a pure capitalist problem systems there are some problems with that, but I think that in design you can think about these longer term financial impacts, and also these more intangible things. [We can ask] how can I bring these in, either as pure quantifications or as some sort of qualitative measures.
Almost all design is an exploration of a designed space, and multiple goals and all of the constraints, and your ideas are just points in the potentially very, very big design space.
If I articulate what are the good and bad things about this one versus that one, then maybe I can make a clear, informed decision. In the process of articulation, you may come up with a third option, which again, does not have all good things and no bad things, but has a different mixture of good and bad things, that on balance you might be better than the two I’m agonizing over.

Ubois:
Is there some way of thinking about responsibility as being part of that calculation?

Twidale:
I think there should be. Everybody in the world where I work talks about concept of the design space [the possible things you might build], but when I look for readings on the concept of design space to explain it to my students there’s precious little there. We use it as a metaphor. All of us who do it know quite well what it means, but there’s a less than clear articulation of it. And I think part of it is that you’ve got a set of multiple goals, and I think that’s the most crucial point, especially when I’m talking to students, because if they come from an engineering world as taught in high school or when you’re undergrad, not real life engineering.
In school, there’s only one right answer and it all turns into nice round numbers so it’s easy to grade, and you’re only trying to optimize one thing. But usually you’re trying to optimize 2 or 3 or 30 things, with multiple goals that can be contradictory, with a whole lot of invidious trade offs. So we talk about these other issues of empowerment, and environmental issues and these other things. Pounds of carbon is a very nice convenient measure, dollars is an even more convenient measure. But even if you can’t use those, as long as you’ve got relative measures, you can say, “well, I don’t know how much it is, but I know this one if way better than that one.” That’s enough for you to start making some decisions. But if I say, “in terms of environmental impact, this option is really, really good, but in terms of empowerment of women, it’s really bad,” should I go for it or not?

Ubois:
Are there some ethical or cultural policies that research institutions could adopt that would encourage that sort of thinking?

Twidale:
I’d be reluctant to say that there should be. I would rather say that these issues should be brought into discussions of design, all design, and we should introduce the concept of design, not just in terms of engineering design, but what you might call social engineering. So you’re designing this program, you know, say for increasing diversity in the work force. I consider that a form of social engineering. And you can say, okay, well we could do this policy or we could do this policy. What are the pros and cons of these policies? Who would gain in this policy? Who would lose? What would be the unintended consequences?
Those design activities should be made explicit and taught to all students, you know, across the curriculum. And that would introduce the discussion of what are your ethical variables. I want to set it out so that we can have a group argument about the relative ethics by you telling me what matters to you. And somebody else saying what matters to them. And so we can figure it out.

Ubois:
Well, to tie it back to IRBs, it seems like there’s a pretty big disconnect there. I mean, ultimately, in the grander sense, so there’s some similar intention, that’s a pretty big disconnect between the IRB process and what you’re talking about here.

Twidale:
Yes.

Ubois:
So I just wonder how you would go about bridging that? Or is it possible? Or do you put the IRBs aside and say this is the new way forward?

Twidale:
I think the trick is to bring these issues to an IRB, and what I’m going to try out on them is getting some kind of more blanket permission, and show them a whole lot of cases that if they don’t give me blanket permission, then pretty much everybody’s breaking the law.

Ubois:
So you find good test cases for this.

Twidale:
Yes, I think so. But I think they find what I’m doing to be sort of madly peripheral, which is quite true, because I want to change the rules of what research is. And we’re not allowed to talk about that.

Ubois:
Can you talk a little bit more about that?

Twidale:
Yeah. So traditionally, research is you plan an experiment very carefully, then you go into your lab, you conduct the experiment, and then experiment is over. Then you analyze the results, and you sort of fit it into a larger theory and publish it. Whereas, what I’m saying is no, I live my research. Every working hour I am conducting research on human subjects. I am subjecting you to the stimulus of my voice and I’m measuring the feedback.
Now IRB can’t cope with that. They say, but that’s what the experiment. Or what they’d like to say is, oh, but that’s not research. And I say, says who? That’s what I do. I mean sometimes they’ve said, oh, you don’t need permission for that, because we don’t think that’s research? I say, thanks very much.

Ubois:
That’s almost an anthropological approach. You know, you go and join the tribe and you have this effect on them and they have an effect on you and that’s expected. Which is a really different ethos …

Twidale:
Yes, I mean, part of what I want to do is sort of invent new ways of doing research that can cope with this strange world we are in. It’s something that you realize that part of the field of research is profoundly uncomfortably with.

Ubois:
One last point – do you have any ideas about improving public dialogue around innovation? There seems to be a consensus that the media doesn’t – isn’t built – to convey nuance about scientific questions. Do you have any thoughts on improving public dialogue?

Twidale:
I’ve, you know, it goes back to I think the one ethical issue I so see a lot in my work and that is the idea of empowerment, and designing better computer interfaces that can convey complex ideas. And that is something which can be incorporated into the public understanding of science, but also it allows people to contribute back. You allow people to say, well, I find this confusing or whatever. It can be fed back to me as a systems designer. So I see that as a research challenge to say, how do we convey something of what it is that we do, and including the fact that we don’t know things.
Like one of the big problems in how science is portrayed is this concept that science knows the right answer. And that’s clearly not the case – I mean even with global warming. Some politicians just want to know “well is it happening or not?” And they can’t cope with the answer, which is that we don’t really know, but we’re 95 percent sure. Some will pause at that point and say “well, I’ll hold out until it’s 100 percent.” Earth will be fried to a crisp before that time. We need to sort of convey the way that research is done and all that’s missing, that uncertainty. There are some things we could do, smaller level things. I think good visualizations of information, could be very powerful.
Have you seen the baby name wizard, that computer visualization of baby names? It’s not exactly profound data, but it’s a very large amount of data, with the popularity of boys and girls names over 100 years, and it compresses truly huge amounts of data into a visualization. Anybody can figure out how to interact with it, and people start asking very sophisticated questions, charting the relative popularity of names against political events. They start imposing things like, Marilyn. Diana, Shirley Temple, people can figure out the socio-political effects of naming. Adolph. And so there is a set of data that is essentially a pile of numbers. You turn it into an interactive graphical computer application and just anyone can use it.
It’s an existence proof that it is possible to design very sophisticated representations of data that your average man and woman on the street can figure out.

Ubois:
Does it convey a feeling of understanding without really conveying understanding? My subjective sense of understanding when I watch television news is probably higher than it should be.

Twidale:
Well, these things can be layered. One TV news program can’t really give you depth, because there’s not enough time. But if there are enough of them, you start to know more about it, and good news reporting tries to do that, and it’s possible with some care.

Ubois:
Thank you very much for taking the time with me.

(read the first part and the second part)