Melanie Mitchell, former student of Douglas Hofstadter and respected AI authority in her own right, recalled attending a conference at the Googleplex in Mountain View, California in 2014, a meeting where her old mentor was presenting to an auditorium full of young and starry-eyed Google AI engineers.
The occasion was apropos: Google had committed, almost 20 years earlier, to taking the lead in the emerging AI industry; and Hofstadter, whose boy-genius masterwork Gödel, Escher, Bach: An Eternal Golden Braid had inspired thousands of young people (including myself) to pursue AI as a career, was the perfect keynote speaker.
Mitchell noted, with some irony, that the Google crowd did not get what they were expecting.
Hofstadter, who had ruminated at length on the nature and possibilities of AI (among other things) in GEB, and who has continued to speak and write about AI in great depth since its publication, presented an unexpected and unsettling counterpoint to the conference’s optimistic theme: that AI has him terrified.
He recounted his experience with David Cope, a musician who had created, in the 1990s, a program that could write complex and emotive music. Hofstadter had predicted in GEB that while he believed a computer would one day write beautiful music, it wouldn’t happen anytime soon. When he met Cope and played a mazurka his program, EMI (for Experiments in Musical Intelligence) had composed in the style of Chopin, he found himself “deeply troubled.”
“Ever since I was a child, music has thrilled me and moved me to the very core,” he told the Google engineers before him. “And every piece that I love feels like it’s a direct message from the emotional heart of the human being who composed it. It feels like it is giving me access to their innermost soul. And it feels like there is nothing more human in the world than the expression of music. Nothing. The idea that pattern manipulation of the most superficial sort can yield things that sound as if they are coming from a human being’s heart is very, very troubling. I was just completely thrown by this.”
Hofstadter further recalled for his audience one of several experiments he had subsequently performed with lecture audiences, one in which he played a recording of an authentic but little-known Chopin mazurka, then played a recording of the imitation Chopin mazurka that had been composed by Cope’s program. He asked for the audience members to vote on which piece was authentic and which was computer-composed. To his shock, most of the audience got it wrong.
Mitchell recalled that Hofstadter paused, as his audience sat in total silence, not sure what to think.
“I was terrified by EMI,” he then continued. “Terrified. I hated it, and was extremely threatened by it. It was threatening to destroy what I most cherished about humanity. I think EMI was the most quintessential example of the fears that I have about artificial intelligence.”
Then he got personal, looking out at the faces of the enthusiastic crowd of eager young engineers and researchers.
“I find it very scary, very troubling, very sad, and I find it terrible, horrifying, bizarre, baffling, bewildering, that people are rushing ahead blindly and deliriously in creating these things.” According to Mitchell, he meant them.
Tens of millions of people share Hofstadter’s terror. They are aware of AI’s rapid, unstoppable rise, about which the media just won’t shut up, and fear the emergence of Skynet-styled horror – that AI will take over the world, and even decimate humanity. The less dramatic are anxious that AI is going to lay waste to the global economy by eliminating hundreds of millions of jobs, erasing entire professions, and one of them will be their own.
But Hofstadter’s terror comes from a different place, Mitchell explained.
“It was not about AI becoming too smart, too invasive, too malicious, or even too useful. Instead, he was terrified that intelligence, creativity, emotions, and maybe even consciousness itself would be too easy to produce – that what he valued most in humanity would end up being nothing more than a ‘bag of tricks,’ that a superficial set of brute-force algorithms could explain the human spirit.”
Later, when they spoke privately, Hofstadter clarified his comment to Mitchell:
“He fears that AI might show us that the human qualities we most value are disappointingly simple to mechanize,” she remembered. “As Hofstadter explained to me after the meeting, here referring to Chopin, Bach, and other paragons of humanity, ‘If such minds of infinite subtlety and complexity and emotional depth could be trivialized by a small chip, it would destroy my sense of what humanity is about.’”
This is an aspect of the changes soon upon us that have gotten no discussion or consideration. Sitting down with ChatGPT for only a few minutes is humbling, even to those who fully understand its inner workings and its limitations: it reminds us that we all possess a layer of passive hubris where the work of the human mind is concerned, as it has been our exclusive domain for all of time – until this moment. If we give the matter the thought Hofstadter clearly has, we quickly realize that AI can, with steely indifference, evaporate that layer of hubris effortlessly.
That the machine can knock us from our mighty perch so casually should terrify us. And equally terrifying, Hofstadter is correct in saying, is that the very engineers and researchers who are hastening AI’s advent with the eagerness of kids awaiting Christmas morning seem utterly clueless about the distressing impact their work will inevitably have on so many who will having nothing to say about it.
It was the right message in the right place at the right time. Whether or not Hofstadter’s message got through to anyone sitting in that auditorium, we may never know. But it got through to Mitchell, who proceeded to write a book addressing the problem he’d called out.
No doubt many in that auditorium would say, even if they sympathized with him, that it doesn’t matter; the genie is out of the bottle. AI is here, and the day when a digital musical master is realized, not just as a one-trick program but as a bona fide superintelligence that can also compose, is inevitable.
“Maybe this is going to happen,” Hofstadter conceded, “but I don’t want it to happen soon. I don’t want my children to be left in the dust.”
Comments