Richard Dawkins, the biologist and author, is complicated. I reached this conclusion in 2005 when I participated in a fellowship for journalists organized by the pro-religion Templeton Foundation. Ten of us spent several weeks at the University of Cambridge listening to 18 scientists and philosophers point out areas where science and religion converge. Alone among the speakers, Dawkins argued, in his usual uncompromising fashion, that science and religion are incompatible. But in his informal interactions with me and other fellows, Dawkins was open-minded and a good listener. Over drinks one evening, a Christian journalist described witnessing an episode of faith healing. Instead of dismissing the story outright, Dawkins pressed for details. He seemed to find the story fascinating. His curiosity, at least for a moment, trumped his skepticism.

I mention this episode because it is illustrative of the thinking on display in Dawkins’s newest book, Science in the Soul: Selected Writings of a Passionate Rationalist. It consists of essays written over the last several decades on, among other things, altruism, group selection, extraterrestrials, punctuated equilibrium, animal suffering, eugenics, essentialism, tortoises, dinosaurs, 9/11, the problem of evil, the internet, his father and Christopher Hitchens. The book showcases Dawkins’s dual talents. He is a ferocious polemicist, a defender of reason and enemy of superstition. He is also an extraordinarily talented explicator and celebrator of biology. He makes complex concepts, like kin selection, pop into focus in a way that imparts a jolt of pleasure. His best writings are suffused with the wide-ranging curiosity that he revealed at the fellowship in Cambridge.

I posed some questions to Dawkins about the importance of scientific reasoning, the problems of AI and consciousness, and President Donald Trump.

[An edited transcript of the interview follows.]

At the Templeton meeting, you described yourself as an agnostic, because you cannot be certain that God does not exist, correct?
This is a semantic matter. Some people define atheism as a positive conviction that there are no gods and agnosticism as allowing for the possibility, however slight. In this sense I am agnostic, as any scientist would be. But only in the same way I am agnostic about leprechauns and fairies. Other people define agnosticism as the belief that the existence of gods is as probable as their nonexistence. In this sense I am certainly not agnostic. Not only are gods unnecessary for explaining anything, they are overwhelmingly improbable. I rather like the phrase of a friend who calls himself a “tooth fairy agnostic”—his belief in gods is as strong as his belief in the tooth fairy. So is mine. We live our lives on the assumption that there are no gods, fairies, hobgoblins, ghosts, zombies, poltergeists or any supernatural entities. Actually, it is not at all clear what supernatural could even mean, other than something which science does not (yet) understand.

In your new book’s introduction you allude to Donald Trump’s election, and say that now “more than ever, reason needs to take center stage.” What would you say to Trump if you had his ear? Do you think you could reason with him?
Mr. Trump, you appear to be laboring under the delusion that you have the necessary qualifications to be president. The manifest failure of almost everything you have attempted during your first six months, coupled with the anarchic chaos that pervades your White House, should give you pause—or would give pause to any person of normal sensitivity.

What advice would I give? Get your news, not from FOX but from all the sources available to a president, many of them not available to the rest of us. Announce your decisions after due consideration and consultation, not impulsively on Twitter. Cultivate common good manners when dealing with people. Do not be misled by the crowds that cheer your boorish rudeness: they are a minority of the American people.

Listen to experts better qualified than you are. Especially scientists. Be guided by evidence and reason, not gut feeling. By far the best way to assess evidence is the scientific method. Indeed, it is the only way if we interpret “scientific” broadly. In particular—since the matter is so urgent and it may already be too late—listen to scientists when they tell you about the looming catastrophe of climate change.

No I don’t think I could reason with Trump. Why would I succeed where so many have failed?

The “reproducibility crisis” in research has raised questions about  science’s reliability. Do scientists deserve some blame for widespread debate over climate change, evolution and vaccines?
It is a real worry, perhaps especially acute in medical research. Part of the problem is the tendency for results to be simplified in order to make a neat, easily summed-up story. And this is exacerbated when recent research results hit the newspapers or other media.

Another problem is the “file drawer effect” whereby papers that fail to disprove the null hypothesis are never published, because authors or editors think they’re too boring. This could theoretically lead to falsehoods being propagated: If enough studies are done, a minority will yield statistical significance even if the null hypothesis is true.

Despite the “reproducibility crisis” there are some scientific conclusions that really are robust and become progressively more so as time goes by. The fact of evolution is one such.

What can be done to resolve the reproducibility crisis?
In the case of the file drawer effect a possible remedy is for all scientists to post on the internet their intention to do an experiment before they do it, and share the results even if these are negative and therefore not appealing to journal editors. On this system, journals would refuse to publish the results of an experiment that was not announced ahead of time.

In the case of scientific findings that hit the headlines too quickly, editors should be less keen on hot, latest news and should give more space to timeless science.

Zooming out a bit, do you share the concerns of billionaire entrepreneur Elon Musk, who has said that artificial intelligence might pose an “existential risk” to humanity?
Elon Musk is a 21st-century genius. You have to listen to what he says. I am philosophically committed to “mechanistic naturalism,” from which follows the conclusion that anything humans can do, machines can in principle do, too. In many cases we already know they can do it better. Whether they can do it better in all cases remains to be seen, but I wouldn’t bet against it. The precautionary principle should lead us to behave as though there is a real danger—a danger we should take immediate steps to forestall. Unless, that is, we think robots could to a better job of running the world than we can. And a better job of being happy and increasing the sum of sentient happiness.

Is consciousness a scientifically tractable problem? Do you favor any current approaches and theories?
It certainly isn’t tractable by me. At times I find myself inspired by the confidence of my friend Daniel Dennett. At other times I lean towards his fellow philosopher Colin McGinn’s pessimism: the view that the human mind is flatly incapable of understanding its own consciousness. Our brains evolved to understand how to survive in a hunter–gatherer way of life on the African savanna—understand the behavior of an extremely narrow range of medium-sized objects travelling at medium velocities. It is therefore a wonder, as [cognitive scientist] Steven Pinker has pointed out, that our brains have advanced to the heights of relativity and quantum mechanics. Maybe this should give us Dennettian confidence. Or maybe the “hard problem” of consciousness is forever beyond us, just as calculus is forever beyond the mentality of a chimpanzee.

What’s your utopia?
My utopia is a world in which beliefs are based on evidence and morality is based on intelligent design—design by intelligent humans (or robots!). Neither beliefs nor morals should be based on gut feelings, or on ancient books, private revelations or priestly traditions.