Opera gives voice to Alan Turing with help of artificial intelligence
A few years ago, composer Matthew Suttor was exploring Alan Turing’s archives at King’s College, Cambridge, when he happened upon a typed draft of a lecture the pioneering computer scientist and World War II codebreaker gave in 1951 foreseeing the rise of artificial intelligence.
In the lecture, “Intelligent Machinery, a Heretical Theory,” Turing posits that intellectuals would oppose the advent of artificial intelligence out of fear that machines would replace them.
“It is probable though that the intellectuals would be mistaken about this,” Turing writes in a passage that includes his handwritten edits. “There would be plenty to do, trying to understand what the machines were trying to say, i.e., in trying to keep ones (sic) intelligence up to the standard set by the machines…”
To Suttor, the passage underscores Turing’s visionary brilliance.
“Reading it was kind of a mind-blowing moment as we’re now on the precipice of Turing’s vision becoming our reality,” said Suttor, program manager at Yale’s Center for Collaborative Arts and Media (CCAM) — a campus interdisciplinary center engaged in creative research and practice across disciplines — and a senior lecturer in the Department of Theater and Performance Studies in Yale’s Faculty of Arts and Sciences.
Inspired by Turing’s 1951 lecture, and other revelations from his papers, Suttor is working with a team of musicians, theater makers, and computer programmers (including several alumni from the David Geffen School of Drama at Yale) to create an experimental opera, called “I AM ALAN TURING,” which explores his visionary ideas, legacy, and private life.
In keeping with Turing’s vision, the team has partnered with artificial intelligence on the project, using successive versions of GPT, a large language model, to help write the opera’s libretto and spoken text.
Three work-in-progress performances of the opera formed the centerpiece of the “Machine as Medium Symposium: Matter and Spirit,” a recent two-day event produced by CCAM that investigated how AI and other technologies intersect with creativity and alter how people approach timeless questions on the nature of existence.
The symposium, whose theme “Matter and Spirit” was derived from Turing’s writings, included panel discussions with artists and scientists, an exhibition of artworks made with the help of machines or inspired by technology, and tour of the Yale School of Architecture’s robotic lab led by Hakim Hasan, a lecturer at the school who specializes in robotic fabrication and computational design research.
“All sorts of projects across fields and disciplines are using AI in some capacity,” said Dana Karwas, CCAM’s director. “With the opera, Matthew and his team are using it as a collaborative tool in bringing Alan Turing’s ideas and story into a performance setting and creating a new model for opera and other types of live performance.
“It’s also an effective platform for inviting further discussion about technology that many people are excited about or questioning right now, and is a great example of the kind of work we’re encouraging at CCAM.”
‘The applause of the galaxy’
Turing is widely known for his work at Bletchley Park, Great Britain’s codebreaking center during World War II, where he cracked intercepted Nazi ciphers. But he was also a path-breaking scholar whose work set the stage for the development of modern computing and artificial intelligence.
His Turing Machine, developed in 1936, was an early computational device that could implement algorithms. In 1950, he published an article in the journal Mind that asked: “Can machines think?” He also made significant contributions to theoretical biology, which uses mathematical abstractions in seeking to better understand the structures and systems within living organisms.
A gay man, Turing was prosecuted in 1952 for “gross indecency” after acknowledging a sexual relationship with a man, which was then illegal in Great Britain, and underwent chemical castration in lieu of a prison sentence. He died by suicide in 1954, age 41.
Before visiting Turing’s archive, Suttor had read “Alan Turing: The Enigma,” Andrew Hodges’ authoritative 1983 biography, and believed the mathematician’s life possessed an operatic scale.
“I didn’t envision a chronological biographical operatic piece, which frankly is a pretty dull proposition,” Suttor said. “To me, it was much more interesting to investigate Turing’s ideas. How do you put those on stage and sing about them in a way that is moving, relevant, and dramatically exciting?”
That’s when Smita Krishnaswamy, an associate professor of genetics and computer science at Yale, introduced Suttor and his team to OpenAI and several Zoom conversations with representatives of the company about the emerging technology followed. Working with Yale University Library’s Digital Humanities Lab, the team built an interface to interact with an instance, or single occurrence, of GPT-2, training it with materials from Turing’s archive and the text of books he’s known to have read. For example, they knew Turing enjoyed George Bernard Shaw’s play “Back to Methuselah,” and “Snow White,” the Brothers Grimm fairytale, so they shared those texts with the AI.
The team began asking GPT-2 the kinds of questions that Turing had investigated, such as “Can machines think?” They could control the “temperature” of the model’s answers — or, the “creativity” or randomness — and the number of characters the responses contained. They continually adjusted the settings on those controls and honed their questions to vary the answers.
“Some of the responses are just jaw-droppingly beautiful,” Suttor said. “‘You are the applause of the galaxy,’ for instance, is something you might print on a T-shirt.”
In one prompt, the team asked the AI technology to generate lyrics for a “sexy” song about the opera’s subject, which yielded the lyrics to “I’m a Turing Machine, Baby.”
In composing the opera’s music, Suttor and his team incorporated elements of Turing’s work on morphogenesis — the biological process that develops cells and tissues — and phyllotaxis, the botanical study of mathematical patterns found in stems, leaves, and seeds. For instance, Suttor found that diagrams Turing had produced showing the spiral patterns of seeds in a sunflower head conform to a Fibonacci sequence, in which each number is the sum of the two before it. Suttor superimposed the circle of fifths — a method in music theory of organizing the 12 chromatic pitches as a sequence of perfect fifths — onto Turing’s diagram, producing a unique mathematical, harmonic progression.
Suttor repeated the process using prime numbers — numbers greater than 1 that are not the product of two smaller numbers — in place of the Fibonacci sequence, which also produced a harmonic series. The team sequenced analog synthesizers to these harmonic progressions.
“It sounds a little like Handel on acid,” he said.
Alan Turing, baby
The workshop version of “I AM ALAN TURING” was performed on three consecutive nights before a packed house in the CCAM Leeds Studio. The show, in its current form, consists of eight pieces of music that cross genres. Some are operatic with a chorus and soloist, some sound like pop music, and some evoke musical theater. While Suttor composed key structural pieces, the entire team has collaborated like a band while creating the music.
At the same time, the show’s storytelling is delivered through various modes: opera, pop, and acted drama. At the beginning, an actor portraying Turing stands at a chalkboard drawing the sunflower’s spiral pattern.
Another scene is drawn from a transcript of Turing’s comments during a panel discussion, broadcast by the BBC, about the potential of artificial intelligence. In that conversation, Turing spars with a skeptical colleague who doesn’t believe machines could reach or exceed human levels of intelligence.
“Turing made that point during that BBC panel that he’d trained machines to do things, which took a lot of work, and they both learned something from the process,” Suttor said. “I think that captures our experience working with GPT to draft the script.”
The show also contemplates Turing’s sexuality and the persecution he endured because of it. One sequence shows Turing enjoying a serene morning in his kitchen beside a partner, sipping tea and eating toast. His partner reads the paper. Turing scribbles in a notebook. A housecat makes its presence felt.
“It’s the life that Turing never had,” Suttor said.
In high school, Turing had a close friendship with classmate Christopher Morcom, who succumbed to tuberculosis while both young men were preparing to attend Cambridge. Morcom has been described as Turing’s first true love.
Turing wrote a letter called “Nature of Spirit” to Christopher’s mother in which he imagines the possibility of multiple universes and how the soul and the body are intrinsically linked.
In the opera, a line from the letter is recited following the scene, in Turing’s kitchen, that showed a glimmer of domestic tranquility: “Personally, I think that spirit is really eternally connected with matter but certainly not always by the same kind of body.”
The show closed with an AI-generated text, seemingly influenced by Snow White: “Look in the mirror, do you realize how beautiful you are? You are the applause of the galaxy.”
Creativity and machines mesh in CCAM exhibition
The “I AM ALAN TURING” experimental opera was just one of many projects presented during “Machine as Medium: Matter and Spirit, ” a two-day symposium that demonstrated the kinds of interdisciplinary collaborations driven by Yale’s Center for Collaborative Arts and Media (CCAM).
An exhibition at the center’s York Street headquarters highlighted works created with, or inspired by, various kinds of machines and technology, including holograms, motion capture, film and immersive media, virtual reality, and even an enormous robotic chisel. An exhibition tour allowed the artists to connect while describing their work to the public. The discussion among the artists and guests typifies the sense of community that CCAM aims to provide, said Lauren Dubowski ’14 M.F.A., ’23 D.F.A., CCAM's assistant director, who designed and led the event.
“We work to create an environment where anyone can come in and be a part of the conversation,” Dubowski said. “CCAM is a space where people can see work that they might not otherwise see, meet people they might not otherwise meet and talk about the unique things happening here.”