Why the humanities matter in the age of AI

For most of the last decade, the advice to students ran in one direction: learn to code, pick a technical major, stay ahead of automation. That advice is aging quickly. As AI grows more capable at technical tasks, the skills that remain valuable are the ones a humanities education is built to develop.

Even AI leaders think so. Daniela Amodei, president of the AI company Anthropic, studied English literature, and she has said that as AI improves, “the things that make us human will become much more important.” Her co-founder Jack Clark, who leads policy at the same company, studied English literature and creative writing. Several of the people building frontier AI were trained to read closely and write well.

What AI does and what it doesn’t

Large language models are good at producing plausible text, summarizing known material, and handling routine analysis. What they don’t do on their own is decide which questions matter, notice when an answer is confident but incorrect, weigh a decision against its human consequences, or explain a difficult trade-off to the person who has to live with it. Those are the moves a humanities education trains directly: reading for implicit claims, testing an argument’s weaknesses, and writing so a reader can act.

The economics have pointed this way for a while. David Deming’s research at Harvard found that nearly all job growth since 1980 has been in occupations that require strong social skills, because that work resists automation. AI has raised the stakes on that finding rather than reversing it.

That adaptability already shows in graduate outcomes. A 2026 LinkedIn ranking published by Forbes found English the most versatile major in the country, with 69% of recent graduates landing jobs in fields farther afield from their major, ahead of business, computer science, and engineering. When the shape of entry-level work keeps shifting, that versatility is an asset.

Employers are starting to recognize it

Hiring data confirms that the humanities matter. A 2026 Strada Institute survey of nearly 1,500 executives and talent leaders found that employers rate critical thinking and communication as the most important skills in early-career hires, and AI literacy as the least. The same survey found more leaders expecting AI to increase entry-level hiring rather than to decrease it by a wide margin. The World Economic Forum, writing in January 2026, framed human skills as “the new advantage” in an AI economy.

There is a meaningful counter-current worth naming. Employer demand for explicit AI skills in entry-level postings has risen sharply, and one NACE analysis found it nearly tripled over a single year. The honest reading is that employers want durable human judgment first, with AI fluency alongside it. A humanities graduate who can do both is unusually well positioned.

The recognition gap

Here is where the story stalls. Employers need these skills, and often say so, yet many hiring managers and recruiters don’t recognize liberal arts majors as evidence of them. A philosophy major who spent three years learning to find the weak point in an argument reads, on paper, as someone who “took philosophy classes.” The skill is real and the demand is real, and the two are separated by vocabulary.

That gap is the reason Unbracket exists. Translation, taking what a humanities graduate has done and naming it in terms a recruiter screens for, is a specific and learnable task. It requires finding the professional word for skills liberal arts degrees cultivate and attaching the evidence that backs them.

The bet

The bet behind this whole project is that the humanities are becoming more valuable at the same moment their graduates are told they are worth less. Closing that gap is partly an economic problem and partly a translation problem. The economics are already moving in the right direction. The translation is the part you can do something about this week.

If you want the practical version, start with what you can do with a humanities degree, or walk through a sample profile.