Are we playing dice with the biosphere?

One of the parallels you bring in is the saga of the nuclear industry -- as an example of a technological innovation where we failed to assess risk properly in the past. The establishment perspective was, there were no problems there; then we had some disasters and the public viewed it as a big problem. Today, though, nuclear power is once again being touted by people like the Wall Street Journal editorial page --

Or Peter Schwartz!

-- as a good option we should be reconsidering. They're arguing that the media and the public overreacted to a handful of incidents and squashed a whole industry.

That's absolutely the argument. Here's where the argument is ridiculous: It's right -- it was an overreaction. It was an overreaction because we were oversold that stuff to start off with. You tell me that nuclear energy is 100 percent safe and too cheap to meter, and you beat on this in every possible conceivable venue you can, and tell people who disagree with you that they're stupid Luddite jerks. It's the same thing in biotech, right? You're an idiot if you disagree. The people who were doing research on anti-nuclear stuff, they had really fringe-y people, too. But they also had people who were making legitimate cases about stuff like, how do you dispose of fuel?

The point for me is, if you want to bring a risky technology into the culture, tell the truth about it to start off with! This is the other main point in the book. Human beings, ordinary human beings like me and you, people who read Salon: You sit us down, and you explain to us what's happening. We have a problem, an energy problem. None of the solutions are perfect. Some are riskier than others. Here are our alternatives. We need to make a decision as a society which one we're going to do. Let's open up this process and make it transparent. Then you have what is truly risk communication, as opposed to propaganda.

These people keep telling us, "no science-based evidence..." Whose science? Whose evidence? I spent three years finding evidence on the other side of this coin that's absolutely peer-reviewed, valid, hard-core research. So it frustrates me.

The same thing happened with DDT, right? DDT is fabulous. DDT ended up being an enormous destructive force in the ecosystem because we decided to use it on everything -- just like we did with antibiotics. And we changed the ecosystem, in some cases irrevocably, and now people are saying, we'd like to bring DDT back. Maybe there are still some mosquitoes somewhere that are susceptible to it. But basically, they're resistant now.

The opportunity to use it effectively is gone.

I feel the same way about genetically modified anything. I don't know whether we have a problem that requires that solution yet. We have a technology. And we have a technology that people want to do something with. And that some people think will be very valuable for human health.

It's eye-opening in "Intervention" when you review how essentially unprofitable, so far, how ineffective, and how really just messy the whole world of biotech is. The field's stereotype is like the picture on the cover of your book: these clean glowing vials of magical elixirs, and tiny little pure bits of gene code getting snipped into other codes, and presto! You get something wonderful.

It is untrue.

Instead, you describe a process with all sorts of surprising baggage, like the "cassettes": the other stuff that gets injected from one organism into another in order to insert the gene you're trying to add. I read that and thought, I'm a journalist and publicly aware person -- I should have known that!

Nobody knows this stuff. I had to figure it all out, too. I was blown away. Compare that to what's on the biotechnology industry's Web site about what biotech is: You take a gene out, and you put a gene in. And the same on the FDA's Web site. They misrepresent the technology to the public, so that we have this false sense of simplicity. And if it's really that simple, it must be safe.

At the end of the book, you introduce the idea of collaborative risk assessment, or "the analytic deliberative process," and describe one example you and Baruch Fischhoff conducted -- a model process assessing the risk of human transplants of organs from genetically modified pigs. Is what you did identical to what's described in the National Academies' "Understanding Risk" report?

We took it a step further. One of the things we realized early in the work on genetic engineering in particular was that we were often talking about events that were very likely of very low probability of occurrence -- but if they did occur, the consequences would be extreme. And generally they were things that had never been seen before. Like, there's never before in nature been a time when a living pig cell rubbed up against a living human cell and had the opportunity to exchange retroviruses. The pig valves that people hear about getting inserted today -- that's dead tissue, those are not live cells.

So you have this never-before-seen situation. For traditional risk analysis, quantitative, there's literally no data. You can try to extrapolate from something that's happened before with greater or lesser degrees of success; you won't know till it's too late. But what we found by talking to a bunch of experts outside of our fields was that scenarios were a good way to think about this stuff.

Now, traditional risk analysts hear the word "scenario" and they say, "You're making up stories!" And a scenario person will look at somebody who's trying to do a quantitative risk analysis on something that's never happened before and say, "You're making up numbers!"

So we tried to figure out a way, and I think we did it, where you could come up with what Baruch called a "computable model." You use scenarios to help you draw a map of the known, the unknown and the unknowable. That gives you a picture of the risk. And as time goes on and you get more data, you can compute more pieces of this model and you can start to come up with actual probability calculations. That's what we did that was different -- we went into the scenario stage. There are lots of ways to do analytic deliberative processes where you don't have to go to those lengths. But the process itself exists. It's been used in lots of very tricky technological situations, and it's very effective.

So what you're doing is pulling together a much broader set of perspectives.

For a traditional risk assessment, somebody would be assigned to do this job, and they would figure out what the model was, and they would go ask whatever experts they thought were important, but it would really be driven by one person -- one paradigm, if you will. In this model you blow the walls off of that. You come into a room; you sit a bunch of people around the table from a bunch of different areas of expertise. The catchphrase is "interested and affected parties" -- experts who would have some relevance to the topic, stakeholders. And you open up the question -- you say, This is the problem on the table. It's purely about making a decision: Should we move forward with this or should we not?

Just by starting there, that's different from saying, we have this new product, can it be approved?

I've been puzzling and trying not to be too cynical for three and a half years now about why this doesn't happen more, when it's so obviously the most effective way to do risk assessment for these kinds of problems. It takes a lot of time and it costs money, and you don't poop a ratio out the other end of it where you can say, OK, we've decided, this is what the risk is. It's much more ambiguous. It's very uncomfortable. People do not like to be uncomfortable about this kind of knowledge.

We need to train our culture to understand that there are things we don't know. And we need to build learning into the process of doing these assessments, so that you don't have to undo a regulation when you find out that something has turned out to be risky. You need a way to accept input while you're moving along. The [analytical deliberative] process makes risk assessment very -- porous.

Next page: "What do you do with the pig carcasses?"