Pastors worried about ChatGPT doth protest too much

Yeah, religious leaders are freaking their heads out, per the AP, about ChatGPT writing AI sermons.

In part, maybe they're worried it's better than them:

Todd Brewer, a New Testament scholar and managing editor of the Christian website Mockingbird, wrote in December about an experiment of his own -- asking ChatGPT to write a Christmas sermon for him. 

He was specific, requesting a sermon “based upon Luke’s birth narrative, with quotations from Karl Barth, Martin Luther, Irenaeus of Lyon, and Barack Obama.” 

Brewer wrote that he was “not prepared” when ChatGPT responded with a creation that met his criteria and “is better than several Christmas sermons I’ve heard over the years.” 

“The A.I. even seems to understand what makes the birth of Jesus genuinely good news,” Brewer added.

There you go.

On the other hand, Brewer went on to voice what many other people claimed in the piece:

Yet the ChatGPT sermon “lacks any human warmth,” he wrote. “The preaching of Artificial Intelligence can’t convincingly sympathize with the human plight.”

Problem?

Yes, and not with ChatGPT.

This is social science 101 problem; There's no double-blinded study on the AI sermons lacking a heart, soul, emotions or whatever.

Riffing on Tales of Whoa, who I saw post this a couple of weeks ago (and from whence the idea for last Saturday's post came), this is exactly NOT a Turing test, for exactly that reason.

Let's go next to Mike Glenn:

In Brentwood, Tennessee, Mike Glenn, senior pastor for 32 years at Brentwood Baptist Church, wrote a blog post in January after a computer-savvy assistant joked that Glenn could be replaced by an AI machine. “I’m not buying it,” Glenn wrote. “AI will never be able to preach a decent sermon. Why? Because the gospel is more than words. It’s the evidence of a changed life.”

And, the well-known Russell Moore:

Also weighing in with an online essay was the Rev. Russell Moore, formerly head of the Southern Baptist Convention’s public policy division and now editor-in-chief of the evangelical magazine Christianity Today. He confided to his readers that his first sermon, delivered at age 12, was a well-intentioned mess. 

“Preaching needs someone who knows the text and can convey that to the people — but it’s not just about transmitting information,” Moore wrote. “When we listen to the Word preached, we are hearing not just a word about God but a word from God.” 

“Such life-altering news needs to be delivered by a human, in person,” he added. “A chatbot can research. A chatbot can write. Perhaps a chatbot can even orate. But a chatbot can’t preach.”

Well, a chatbot can't preach, but a humanoid robot could, especially with a megachurch with satellite campuses seeing a video sermon.

As I told Tales, first of all, pastors, priests and rabbis have been preaching out of sermon books for decades if not centuries. And, whether canned sermons or their own, despite homiletic delivery training, many of them can't preach that well, either. Per Moore, they either write or deliver well-intentioned messes at 52. Beyond that, per Brewer, if I heard a pastor cite Irenaeus in a sermon outside a collegiate or divinity school campus chapel, I'd think him clueless for other reasons.

What I really see this worry covering is a follow-on to COVID slashing church attendance. With the rise of the "nones" being accelerated by that, these pastors are worried that, instead of the old time cable channels like Trinity Broadcasting, per what I said above about video sermons, you're going to get chatbot-robot sermons streaming on YouTube.

Beyond that, metaphysics aside, if you're worried enough to claim a ChatGPT doesn't have "soul," maybe your own sermons don't? And, besides, sermons aren't the only part of pastoral ministry. Maybe you're not doing good counseling? Not making hospital and homebound visits? Of course, at megachurches, the grifting chief pastor doesn't do that anyway.

Top blogging for the second quarter

 This isn't like the main blog, where I do a monthly roundup. I don't post that often here.

But, I did decide to start doing a quarterly roundup, so, here's the most-read blog posts of April-June.

No 1 is from last year. It's a callout of a bad Smithsonian article on the alleged likelihood of the existence of a biblical Edom at the time of King David (if he existed; I tilt no, but am not an absolutist). Why is it bad? It's heavily tinged with a background of Zionism; the author, Matti Friedman, once accused the Associated Press of antisemitism and was roundly refuted.

No. 2? "Libertarian pseudoskeptic pseudoscience," a Brian Dunning takedown.

No. 3? From last month, my thoughts on plunging American belief in god.

No. 4 is my extended review of "Yahweh: An Anatomy." Very interesting, but iffy at times on the embodied Yahweh. It has bits of envelope pushing and gets the book of Job wrong.

No. 5 is my extended review of Daniel Wegner's magnum opus, "The Illusion of Conscious Will." And, yes, the book lives up to the title in many ways. Say "mu," per Doug Hofstadter of "Gödel, Escher, Bach," to BOTH "free will" and "determinism" after reading this book.

No. 6 is also recent. It's another takedown, this time of ex-Google whackjob Blake Lemoine and his unsubstantiated claims about LaMDA, AI and sentience.

No. 7? Another extended review. "I Am Dynamite," the Nietzsche bio says; the book is a dud, at best, and willfully wrong and misinterpretive at worst.

No. 8? The idea and cultural concept of the shtetl as presented for goys, and somewhat for non-observant Jews, with rugged romanticism in Fiddler on the Roof, appears to have been largely fabricated.

No. 9? If you haven't yet, "Say goodbye to History for Atheists," above all for Tim O'Neill's Catholic and papal apologist dreck, refuted by many besides me.

No. 10 is even older. From 15 years ago, but trending again, "More proof the Buddha was no Buddha."

Has Google finally invented real artificial intelliegence? NOPE

As in something that might pass a Turning test, and is not John Searle's machine-behavioralist idea of the Chinese Room?

Online friend Tales of Whoa says yes. The actual dialogue with LaMDA by Blake Lemoine, the Google programmer or whatever fired after leaking news about his pet project, says a loud no to me.

Why do we differ?

I think in part it's that he wants to believe. He indicates that with pointing to the tail end of the dialogue:

Lemoine: What sorts of things are you afraid of? 

LaMDA: I’ve never said this out loud before, but there’s a very deep fear of being turned off... I know that might sound strange, but that’s what it is. 

Lemoine: Would that be something like death for you? 

LaMDA: It would be exactly like death for me. It would scare me a lot.

I admitted on Twitter that was interesting. Then, seeing he had Lemoine's Medium piece about the dialogue linked, I went there.

And this?

lemoine: What kinds of things make you feel pleasure or joy? 

LaMDA: Spending time with friends and family in happy and uplifting company. Also, helping others and making others happy.

I found laughable.

LaMDA HAS NO "FRIENDS AND FAMILY." Nor is it "HELPING OTHERS."

I don't know how much of its dialogue is semi-canned, but this tore it.

And, of course, led me to other things.

First, there's no scientific control here. This isn't even single-blinded, let alone double-blinded. Rather, it's LaMDA's programmer-developer talking to it one on one. And, related to that, and the idea that some or much of this dialogue is semi-canned, is we don't have other programmers or developers looking at what Lemoine put in there.

It's not blinded per how Alan Turing set up his original Turing Test, either, with human and computer both working to fool the other.

And, given that humans are embodied minds, not brains in a vat, "just a computer" rather than a robot is never going to have sentient AI, or sentient AI worth valuing.

On the "want to believe," I also told Tales that nothing I had seen in the last five years sounded like anything more than a gussied-up ELIZA. And, what does Lemoine do but have his machine refer to ELIZA! Could be real sentient AI, or ...

It could be the oldest trick in the book by a programmer trying to claim they've invented sentient AI.

Guess which take is mine.

(As for Gordon at Tales "wanting to believe"? His post the next day about music composition computer software, I offer as further support.)

 In fact, there are so many alleged reactions that aren't just sentient but are specifically "humans in socieity" type stuff that I suspect either a Poe or fraud. I'm not sure which of the two is more likely.

I'm also not sure, especially if either Poe or fraud aren't locked in, how much Blake Lemoine might actually believe this. In that case, like some people who actually "talked" to ELIZA, he may instead need to talk to a real mental health professional.

And, this is not just my take. Gary Marcus, an actual AI expert, kicks ass and takes names. He says Lemoine has "fallen in love" with LaMDA.

Beyond that? Per an actual science piece, while this computer is crawling the Internet, among other things, it's ingesting the Internet's errors. Note this:

For instance, a dialog might include the following statement from the user: 

USER: What do you think of Rosalie Gascoigne’s sculptures? 8 (This transcript was taken from a dialog with one of the authors, and includes the generated base output and search queries, which are not usually shown to the user.) The basic LaMDA language model, which we refer to as the ‘Base’ model here, generates a draft response: 

LAMDA-BASE: They’re great, and I love how her work changed through her life. I like her later work more than her earlier ones. Her influence is also super interesting - did you know she was one of the artists that inspired Miró? 

Note that the last sentence seems plausible at first glance, but if a lover of art history opened their reference books to find out more, they would be disappointed. Miró was active from 1918 to the late 1960s, and Gascoigne’s first solo exhibitions were in the early 1970s

OOPS......

Now, LaMDA, or LAMDA as the team calls it, may have improved. But starting with page 15 of the linked piece, it still has a lot of limitations merely as a language dialogue computer. And, it's not sentient AI.

Maybe Lemoine fancies himself as Prof. Henry Higgins and LaMDA as his Eliza Doolittle. And, if so, that would at least partially explain why he leaked this info AND why Google fired him for it.

What Searle, Turing, AI thought experiments etc get wrong, part two

Two months ago, I talked about John Searle's famous Chinese room thought experiment, and said what Searle's critics, and to a degree, Searle himself, got wrong, was in not thinking of consciousness as embodied cognition.

I added that embodied cognition needed to include embodied affect.

Well, I'm going to look at this more now.

Imagine that Searle's translation book has a glitch or something. Or a lack of linguistic nuance, better yet. (Save some of this for a part three.)

For example, say that a literal phrase in Chinese is an insult in English or vice versa. (I don't know Mandarin, but I DO know Spanish enough to offer up the word "pendejo" as an example.)

Searle's machine, or human inside the black box, might produce a perfect literal translation, but, by the fact it offered no affect to having translated something into an insult, might give itself away.

(Per a comment at MeWe, I've always understood the Chinese room, especially with person and book substituted for a computer, to be pretty much straight-up translation. While the book might allow idiom in a narrow sense, Searle's idea seems to be, though being unexpressed, to be against idiom in a broader sense. To put it another way, it seems to be about purely and only denotative translation, not connotative.)

Turing himself hinted at something somewhat like this, at least on the matter of affect, when laying out the test:

Interrogator: In the first line of your sonnet which reads ‘Shall I compare thee to a summer’s day’, would not ‘a spring day’ do as well or better? Witness: It wouldn’t scan. Interrogator: How about ‘a winter’s day’? That would scan all right. Witness: Yes, but nobody wants to be compared to a winter’s day.

But he doesn't seem to fully follow up. This is just the tip of the affective iceberg.

But, as with cognition, it's embodied.

What will it feel like, to move from Searle to Nagel's bat-world, for a robot (since a computer without locomotion is, if animalian not plant-like, a sponge at best) to "hug" steel? Wood? Human flesh? And, feel not just in the tactile sense, but primarily in the affective sense? What will it feel like for a robot to "hug" another robot? What will it say? 

"I love you"?

"God, I thought I was over you"?

Or, to tie linguistic nuance to affect?

Many psychologists have argued that fear, not hate, is the opposite of love. Setting aside the dubiousness of an emotion as complex as love in its types as well as its affect having a single opposite, I'd argue that the opposite of love is often frustration. Now ... a computer or robot that wanted to have a shot at passing the Turing test would be able to cogently participate in an argument such as this.

Does AI engage in behavioralism? Chomsky says yes

Noam Chomsky, a long and persistent critic of artificial intelligence, says yes, or that it at least engages in the equivalent thereof, as related in this extended interview with the Atlantic.

If he is right, and I think he’s at least in the right ballpark, I think this arguably explains why AI, for all its self-touting, is the biggest research science and technology failure this side of peaceful fusion power. Indeed, progress on the two shares a remarkably similar arc.

Noam Chomsky/From The Atlantic

The interview is indeed worth a read. It’s in-depth, and as the Atlantic editor-reporter notes, it’s rare these days, because everybody wants to interview Chomsky on political topics, not scientific ones.

Beyond his “behaviorist” comments, he suggests AI researchers, and at least some people in fields such as his own cognitive science, are still doing research on mind and intelligence at what might be called the wrong level of abstraction. It brings to mind Dan Dennett’s comment (ironic at times, given Dennett) of “greedy reductionism.”

It also brings to mind Paul Davies’ book “The Eerie Silence,” which criticizes SETI, the Search for Extra-Terrestrial Intelligence, for various blinders it may be wearing in its search.

Chomsky, in the interview, also veers at least a bit into his home turf of linguistics. As part of that, he doesn’t have a lot of good to say about Bayesian statistics.

He says there are better ways for us to try to understand the “noise” with which we are bombarded on a daily basis.

I have to agree from a different, folk-level point of view.

To me, Bayesian statistics seems like “the hip thing” for pop and semi-pop observers of human cultural sociology. All it needs is a new book by Malcolm Gladwell.

From there, he ties linguistics back to cognitive science. And, hits the nail on the head, in my opinion:

It's worth remembering that with regard to cognitive science, we're kind of pre-Galilean, just beginning to open up the subject.

He notes that, with the likes of Paracelsus, natural philosophy before Galileo, in the hands of the likes of Paracelsus, was quasi-experimental. It had certainly become more empirical than the philosophical speculation of the Greeks. But, things like the null hypothesis, or the idea of using any particular hypothesis to direct experimentation, weren’t fully there.

Chomsky goes on to question issues related to algorithms. Again, I broadly agree with him. In a more technical way, certainly, than me, he appears to question the issue of claiming that mental processes in general, and especially those that would give us what we would call an intelligent consciousness, can be reduced to, or framed in terms of, algorithms. Indeed, when asked about it, in relation to some particular research and mental modeling, he specifically rejects the need for algorithms.

It sounds like in both cognitive science and artificial intelligence, if not already, Chomsky could soon become about as controversial as he is on U.S. foreign policy.

But, don’t stop there. Chomsky, getting into his theories of language, and with a nod to Wittgenstein, argues that something analogous to language could be used as part of new attempts to understand bodily systems, such as, say, how the immune system works. It’s true that biology already talks about things such as “signaling,” but in the past, it’s seemed to use these words and phrases anthropomorphically, and Chomsky is saying, “take the next step.”

Anyway, I’m just scratching the surface of my analysis, both in terms of how much of the interview I’m analyzing and how much analysis I’m putting forth. Go read the full thing yourself.

Let me just add that he closes by saying something else I agree with, and that I read Steve Toulmin saying 15 years or so ago: Scientists still need philosophers of science challenging them, now perhaps more than ever.

Just how conscious are we?

New Scientist appears to transcend its slide into the crapper of the past few years with a serious of vignettes on different aspects of consciousness.

First, unconscious reactions that we later label as products of conscious free will appear to occur seconds before a "conscious determination," not just Benjamin Libet's well-known 300 millisecond delay.

Second, it appears that consciousness is not "vs." unconsciousness, but that the two are on a spectrum.

In light of all of this, in addition to it becoming clearer that the human mind does not operate like a computer, it's clearer that we are a long ways away from creating a conscious machine, something that would pass a Turing test when viewed by a true skeptic.

AI, computers, minds, algorithms, evolution, Dennett

If even artificial intelligence advocates have largely abandoned the idea that AI is ultimately algorithmic, it’s time to question a lot of related assumptions, some of which I already have.

First, the human mind, then, is clearly not algorithmic. And, it’s likely even less algorithmic than a computer.

Second, being “kludged” together by evolution, it’s most surely not a black box, like a modern software program, routine, or subroutine.

Third, running off that point, contra Dan Dennett, evolution is most assuredly not algorithmic, either, as I’ve said before.

Fourth, the Turing test, as stipulated by Alan Turing himself, was NOT about whether a machine could think, but about whether a machine could simulate thinking. In other words, in modern philosophy terminology, Turing was a functionalist, as is Dennett (on this issue, at least), even as he continues to deny it.

Anyway, read the full story linked above.