College graduates with liberal arts degrees have long wandered in the desert of obscurity, saddled with society’s low expectations. In a world that calculates value based on return on investment, a humanities degree is nothing more than an underperforming investment.
And what’s to become of freshly minted Bachelor of Arts recipients?
English major? I hear Barnes & Noble is hiring.
Political science degree? What’s next—law school?
Master’s in philosophy? Lemme have an iced grande mocha, two pumps.
Meanwhile, future accountants and programmers use spring break to sort through all their job offers. Is it possible that this condition will flip thanks to an unexpected technological breakthrough?
Then came artificial intelligence.
The overnight AI sensation took a while to get here. Dr. John McCarthy coined the term as part of the Dartmouth Conference in 1956, where all the biggest left-brain thinkers got together to talk about machine intelligence. We’ve seen many mathematical milestones along the way and a few developments that might interest a right-brained reader. In the 1990s, computers were trained in chess and managed to defeat a world champion. It turns out that chess is easy if you can evaluate 200 million moves per second. By 2011, computers trained in natural language models defeated Jeopardy! champion Ken Jennings, thus signaling robotic domination of civilization.
Or was there a new hope? Could liberal arts majors unite and strike back?
Liberal Artists vs. AI
In November 2022, OpenAI launched ChatGPT. It became one of the world’s hottest apps, gaining one million users in five days. Before it was two months old, the conversational AI-powered chatbot earned 100 million active users . . . including a few liberal artists.
We spontaneously tested the limits of chatbots, giving them word problems, riddles, and history quizzes. Shockingly, chatbots often responded with incorrect answers. At times, they made up names and events. We asked AI to generate a Seinfeld plot. The results confirmed that a stand-up comedian is smarter and infinitely more entertaining than AI.
AI needs a partner who can navigate the semantics during AI development, prompt design, and result validation.
The unintended consequence of the rollout was that we all learned all at once that AI was not magic. At times, it hallucinated (generating inaccurate answers) or confabulated (filling in missing information). It even demonstrated algorithmic prejudice (a bias that introduces skewed or discriminatory data sets).
AI: Master of Syntax, But Semantic Fraud
Hey chatbot, define syntax and semantics.
Chatbot: Syntax means the rules of a language, letters, symbols, and words; semantics is the meaning of the words.
So, AI is 99.99% accurate when it comes to the rules of language, but it’s hit-and-miss when it comes to meaning. Before we look at the role that teachers, historians, and baristas may play in the AI revolution, it may be helpful to briefly look under the hood of how AI generates its output.
I Use AI, But Don’t Ask Me to Explain AI
Hey chatbot, how does generative AI work?
Chatbot: We play the next-word game. Think of us as probability machines. You give a prompt and based on our training, we calculate the highest probability next word.
Since liberal artists may be queasy with all this talk of math, let’s have an example.
Hey chatbot, if I say, “My drink is sour. I better add BLANK.” Solve for blank and answer in the form of a question.
Chatbot: There is an 85% probability that the right answer is sugar, a 5% probability that the answer is water, a 2% chance the answer is honey. Let’s go with: What is sugar?
Based on what it does and how it does it, perhaps we shouldn’t call it artificial intelligence. If more liberal artists had been invited to that Dartmouth meeting, maybe we’d call the generation of predicted content something colorful like Speculum (Latin for “mirror”); or a descriptive name like Word and Image Machine (WIM); or maybe a name with alliteration: the CopyCatter (CC).
Since AI is clearly a math and probability machine, it may seem like bad news for liberal artists. But no. We are needed to maximize its potential. In fact, AI may pose a greater threat to our accounting and programming colleagues, as their work largely involves constructing the optimum solution based on predefined rules. They live in a syntax-heavy world.
Liberal Artists and AI
AI needs a partner who can navigate the semantics during AI development, prompt design, and result validation. Philosophers are finding jobs to help train AI models. English majors are born prompt engineers. Social science students are naturals for detecting bias in AI models. Political science majors who are putting off law school are experts at giving AI output a skeptical evaluation. In fairness to AI chatbots, they are getting better every day—likely in part to the contribution of liberal artists who are helping them develop better answers.
More broadly, research from Harvard Business School demonstrates that the highest performing teams use AI as a collaborator: “AI serves not just as an information provider but as an effective boundary-spanning mechanism, helping professionals reason across traditional domain boundaries and approach problems more holistically.” The research goes on to suggest that AI helps elevate the performance of less-experienced team members. And we liberal artists are grateful.
Without the burden or backing of research, only informed by my twenty-five years as a Big Four consultant, I believe organizations want to hire people who are strong communicators, problem solvers, and eager learners. And maybe the business world is beginning to recognize the requirement for team members with liberal arts training to balance the syntax-heavy results generated by artificial intelligence.
“That which we call a rose, by any other name, would smell as sweet.”
Who better to construct a thoughtful prompt or conduct the smell test on AI results than a team member who studied history, philosophy, or literature? Of course, it is not fair to compare AI’s output to the product of playwrights. William Shakespeare had the advantage of experiencing love and the scent of a rose. AI can only repeat what it’s been told.