What Do the Turing Test and ID Have in Common?George D. Montañez shows that if a test can detect intelligence in computers, a test could also detect intelligent design in nature
Walter Bradley Center director Robert J. Marks and Harvey Mudd College computer science prof George D. Montañez continued their discussion from last week, Can machines think? (partial transcript) in this week’s podcast, taking in Elon Musk, the universe as a computer sim, and the Search for Extraterrestrial Intelligence (SETI) along the way:
The Turing test, artificial intelligence, and intelligent design
03:43 | Origins: Why some thinkers have held that life was planted on Earth by aliens
Robert J. Marks: There seem to be, at least to my knowledge, three different flavors of intelligent design. One, of course, is from the Christian perspective, that believes that God is the creator, Another one is directed panspermia where the complexity we see on Earth is too complex to understand or to evolve and therefore must have been planted here by aliens. .And panspermia is just the general way of doing that. In other words, life could have come here on a meteor. Directed panspermia means that there was an alien life form that actually came and planted it here.
DNA double helix discoverer Francis Crick (1916–2004) and iconic astronomer Fred Hoyle (1915–2001) were both atheists who accepted the idea of panspermia as the origin of life on Earth.
04:25 | Elon Musk: Are we living in a simulation?
Robert J. Marks: The third one is, I would propose, is due to Elon Musk. He says that things are so complicated here that we must all be simulations—we can’t have been generated by chance so we must all be simulations by some superbeing.
Well-known public figures in the sciences have toyed with the idea that we live in a vast computer sim, including astronomer Martin Rees and MIT PlayLabs founder Rizwan Virk take the idea seriously as well.
04:44 | The Turing test, artificial intelligence, and intelligent design
We’ve talked about the Turing test [for detecting whether a computer has become an intelligent agent] and artificial intelligence and intelligent design. Your paper makes a connection. Could you elaborate on that?
George D. Montañez: I came to a realization when thinking kind of abstractly about what the Turing test is doing. So essentially the Turing test is looking at artifacts. You have words on a computer screen and you have a person who, based on these artifacts is trying to determine whether the cause of them was an intelligent agent or was some sort of unintelligent process. And the Turing test presupposes that, based on the artifacts and the observations alone, you should be able to make that determination.
So for the Turing test to be passed, we want to be able to say that the system is actually intelligent. And for the Turing test to be meaningful, you need there to be some sort of difference between these causal powers and for these differences to be empirically observable, empirically verifiable. And I realized that, if you can do that, this is a major core assumption not just of the Turing test but also of intelligent design detection, which is that we’re supposed to be able to look at artifacts and determine from those artifacts whether or not the cause was an intelligent agent or some unintelligent process.
So it’s kind of this interesting tension because a lot of materialist reductionists would say, especially if they’re believers in strong AI, that perhaps the Turing test is the scientifically valid test of whether there was an intelligent cause or not but when this is applied to things like biology, there can be no scientifically valid test for whether or not the cause of a system was intelligent.
And so the paper draws out the parallels between these two methodologies and shows that they kind of rise and fall together so you can’t have your cake and eat it too, essentially. And so you’re faced with a choice. You can say that the Turing test is not a scientifically valid test and neither is intelligent design theory or if you say that the Turing test is scientifically valid, you cannot a priori rule out methods of design detection such as those used by intelligent design theorists.
07:11 | Search for Extraterrestrial Intelligence (SETI)
Robert J. Marks: Search for Extraterrestrial Intelligence (SETI) … isn’t that kind of a Turing test in a way? A variation of it, I would say.
George D. Montañez: Yes, in the sense that Search for Extraterrestrial Intelligence (SETI is looking at artifacts which are radio signals primarily and trying to determine form observations alone whether the cause was an intelligence or something else that was unintelligent. To the degree that Search for Extraterrestrial Intelligence (SETI could actually work in the real world, it is based on that same set of core assumptions.
Essentially, Montañez shows that the materialist must choose between a Turing test that can detect truly intelligent computers and valid intelligent design hypotheses. Or he can choose no Turing test and no valid intelligent design hypotheses. But he can’t have both. Of course, he could also choose not to be a materialist and face a different array of choices.
It turns out that every belief about the nature of things comes with a cost. There is, as design theorist Bill Dembski says, no free lunch.
Here’s the abstract of the paper.
“Can machines think?” When faced with this “meaningless” question, Alan Turing suggested we ask a different, more precise question: can a machine reliably fool a human interviewer into believing the machine is human? To answer this question, Turing outlined what came to be known as the Turing Test for artificial intelligence, namely, an imitation game where machines and humans interacted from remote locations and human judges had to distinguish between the human and machine participants. According to the test, machines that consistently fool human judges are to be viewed as intelligent. While popular culture champions the Turing Test as a scientific procedure for detecting artificial intelligence, doing so raises significant issues. First, a simple argument establishes the equivalence of the Turing Test to intelligent design methodology in several fundamental respects. Constructed with similar goals, shared assumptions and identical observational models, both projects attempt to detect intelligent agents through the examination of generated artifacts of uncertain origin. Second, if the Turing Test rests on scientifically defensible assumptions then design inferences become possible and cannot, in general, be wholly unscientific. Third, if passing the Turing Test reliably indicates intelligence, this implies the likely existence of a designing intelligence in nature. – Montañez, George. (2016). Detecting Intelligence – The Turing Test and Other Design Detection Methodologies. 517-523. 10.5220/0005823705170523.
Read some excerpts from Can machines think? at How you can REALLY know if you are talking to a computer Claims that a given program has “passed the Turing test” should be treated skeptically because a program can be optimized to pass the Turing test without demonstrating any particular intelligence at all.
Also, by George D. Montañez: AI: Think about ethics before trouble begins. A machine learning specialist reflects on Micah 6:8 as a guide to developing ethics for the rapidly growing profession (George Montañez)
On the universe as a computer sim: How do we know that our universe is not a sim world? It’s an interesting idea, say Bradley fellows, but for a number of reasons, it is not credible