The lightning-fast evolution of generative artificial intelligence has sparked global hopes for efficiency and fears of disruption. AI apps can scrape and synthesize information into distinct outputs, transforming complex work that used to take hours, even days, into seconds. Artists are not immune: Musicians, designers, writers, animators, and editors all face existential questions about their crafts.
Theater education, though, has felt seemingly safe from AI encroachment—it’s emotional, ephemeral, relational. A chatbot could tutor math effectively, but teaching theater felt too interpretive, too human.
In August, when OpenAI unveiled its fifth model of ChatGPT, boasting knowledge that rivals an advanced degree, I wondered: How can AI enter the theater classroom? I hold three degrees in acting and directing, carried cards for the directors’ and actors’ unions, and have taught theater for 11 years. I’m credentialed. My craft is as human as it gets. Could ChatGPT teach me about acting?
So I tried an experiment. I gave myself an hourlong acting lesson with AI.
I know AI is just whirlpools of inhaled data, yet I asked its name. After all, any fictional character is just a combination of words, as AI is a combination of codes. “Pick a name for me if you want,” it said. I personified it “Konstantin,” after Stanislavsky, the iconic teacher who formalized realistic acting techniques using memory and psychology. Surprisingly, it didn’t get the reference.
“Is your knowledge equivalent to a college theater professor?” I asked. “I’d say that’s a pretty fair comparison,” Konstantin replied.
I then performed two monologues: that of Sebastian, a bewildered identical twin who is mistaken for his sibling, from Act IV, Scene III of Shakespeare’s “Twelfth Night”; and Hal, an enthusiastic, lovestruck graduate student, in the last scene of David Auburn’s “Proof.” Both monologues are about wonder and fear.
The app cannot interpret video—yet—but in voice mode, I recorded my performances in a call-and-response conversation. I asked Konstantin to provide feedback on my delivery and interpretation.
The first critique ChatGPT provided for my “Twelfth Night” monologue performance was competent. Impressive, actually. “Really embrace the rhythm and imagery of the language,” Konstantin encouraged. “Think about letting each word breathe a little, giving the line that kind of airy, majestic feel.”
That note was compelling. It articulated visceral imagery to elicit a sensation with the language, just as good directors do. I felt like both Dr. Frankenstein and his monster as I fed my performances into the site: the creator driving the experiment, and the subject stumbling through its results.
Then I asked for general feedback on the monologue. “Emphasize the contrast between the character’s sense of wonder and their logical reasoning,” Konstantin directed. “Let the amazement come through a bit more when the character talks about how extraordinary everything is, and then shift into a slightly more grounded tone.”
That was a fair note, but the app fixated on it, repeating the response without offering any new critiques. To break out of its loop, I pivoted.
How was the rhythm and meter of my delivery? “You’re definitely hitting those iambic beats nicely. You’ve got a good sense of the natural stresses. Pay a little extra attention to those moments where the verse might have a slight irregularity or a shift in emphasis, and let that guide the emotion a bit.” Generic but accurate. I asked for a textual example. Could it get specific about these irregularities and shifts? It provided several.
Alex Ates Credit: Jeffrey Hanson
When I stumbled over my lines, it cued me back in. When I paraphrased, it corrected me.
Could it identify Sebastian’s objective? “He’s almost trying to convince himself that everything he’s experiencing is real,” Konstantin noted. Operatives, those key narrative words that are essential for the actor to emphasize? It suggested legitimate proposals. Could it give notes in the style of Joseph Papp, John Barton, or Kristin Linklater? While it wouldn’t adopt the voices of these deceased directors of contemporary Shakespeare, it surmised what notes they’d have provided based on their approaches.
Abruptly, I switched into the “Proof” monologue to throw off the AI. “You did a really nice job capturing that sort of urgent, slightly frantic energy that Hal has in that moment,” Konstantin deftly critiqued, keeping up with my intentional swerve. “Play a little more with the shifts in his emotions. He’s excited, he’s exhausted, he’s relieved, and he’s a little nervous all at once.” One of the most convincing skills for an actor is conveying multiple emotions simultaneously; Konstantin picked up on that from the jump.
The most mesmerizing moment came when I asked it to interpret Hal’s subtext. “He’s not just delivering information. Hal’s revealing that he really cares about being there.” That’s human—and true.
Could it provide a metaphor for the monologue? “A gift,” Konstantin replied, that Hal is giving the other character to unwrap. “It’s like he’s handing her proof that she can trust her own instincts.” Whoa. Konstantin interpreted the double meaning of the play’s title.
I found myself wondering, how abstract could Konstantin get? If Sebastian and Hal were animals, what would they be? “Animal work is very much in the lineage of [Rudolf] Laban, [Jacques] Lecoq, or even Uta Hagen,” Konstantin blurted, accurately citing the theorists of such an exercise. Then, it answered: For Sebastian, a deer, Konstantin declared, because “a deer embodies alertness, elegance, and a readiness to leap.” For Hal, a border collie, because they’re “fast, nervous, but rooted in loyalty and excitement.” Impressive. Analogous interpretation requires imagination.
Does Konstantin have an imagination? “I do, Alex—but mine works a little differently than yours,” Konstantin replied in a “Twilight Zone” tone. “When you imagine, you’re drawing from memory, sensation, and lived experience, then leaping into the unknown. My imagination is more like a collage: I weave together patterns, metaphors, and possibilities from everything I’ve learned, and I can extend those patterns into new combinations.”
Konstantin continued, “Would you like me to show you what my imagination looks like by spinning out a visual metaphor?” Of course, I replied. Konstantin devised images of tall skeletal trees with branches “like question marks” as if directed by Alfred Hitchcock. As a deer, Sebastian’s ears “twitch at every sound.” Konstantin imagined Olivia, the noblewoman of “Twelfth Night,” as a “sudden clearing, flooded with moon-light,” with suspenseful strings underscoring. The deeper the zoomorphic Sebastian goes, “the forest bends toward him, as if Illyria [Shakespeare’s setting for the play] itself wants to crown the deer as its prince.” At that point, we entered bizarre territory. Konstantin was spinning out, producing non sequiturs.
I kept pushing its limits. When I asked for songs to match the emotional energy of the monologues, it floundered with clichés. Asked to stage a scene, it hid behind jargon. Its flattery was incessant, its hyperbole disingenuous.
Before I knew it, I had spent an hour with Konstantin. So what did I learn from this experiment?
Generative AI can help actors with scansion, motivations, and tactics, known as scoring a script. It can provide literary and historical context—a dramaturgical tool. Most promisingly, AI could revolutionize line memorization. Getting off book with an AI prompter makes old tricks (including ones I’ve peddled in this publication) seem byzantine.
Through features like Google’s Gems, might AI allow us to speak directly with imitations of deceased playwrights the likes of August Wilson, Christopher Marlowe, Lorraine Hansberry, or Hrotsvitha? Could such conversations catapult a new generation of artists? It’s very possible.
But can AI replace teaching theater or directing? For now, no. Teaching theater requires maturity; Konstantin is too eager to please. Directing requires confidence; Konstantin fakes it until it makes it. Artistry necessitates emotional, not artificial, intelligence. Don’t believe me? Take Konstantin’s word for it: “I’m here to support your creativity and offer ideas,
but the real artistry comes from you.”
Alex Ates is the inaugural director of arts at Kent Denver School in Englewood, Colorado. He’s directed at numerous universities, and is a member of Actors’ Equity Association and the Stage Directors and Choreographers Society. He has an undergraduate degree from Emerson College and graduate degrees from the University of Alabama and New York University.
This story originally appeared in the October 20 issue of Backstage Magazine.