- The editorial director of NotebookLM told BI that humanities skills are gaining value in AI.
- Steven Johnson says a new “AI wrangler” role requires knowledge of models and what they can do.
- He said philosophical and psychological skills are increasingly important.
Traditionally, a degree in the humanities wouldn’t translate to a job in tech — but that might change in the age of AI.
In an interview with Business Insider, Steven Johnson, who co-founded several startups and worked as an author before becoming editorial director of NotebookLM at Google Labs, said humanities might be making a comeback.
Johnson, who is helping to build Google’s AI-powered note-taking and research tool NotebookLM, said in the age of large language models, there’s a “revenge of the humanities.” Not only is an English major valuable, but philosophical and psychological skills are also useful.
“There’s just a whole set of questions around AI that no one was thinking about, except for philosophers, until about two years ago,” Johnson said, adding that “those kinds of philosophical skills are really important.”
Johnson said fine-tuning a model’s tone and conversational mode is a “huge part” of the product. He pointed to Amanda Askell, a philosopher who works as an alignment fine-tuning researcher for Anthropic, as an example of someone who does “character training” for AI models.
In a June interview posted by Anthropic on YouTube, Askell said philosophy doesn’t always line up with every aspect of her work, but building Claude’s character feels “philosophically rich.” She said she has to work through complex questions like whether AI models should have moral considerations and what human principles they should be trained on.
“Lots of people are like, ‘See, I told you the degree would be useful,'” Askell said in response to a question about whether it was strange to be a philosopher training an AI model.
Google and Anthropic aren’t the only companies recognizing the value of a humanities degree. One AI startup founder told BI they seek out and said those with liberal arts backgrounds have a “creative and human-centric approach” and a strong understanding of how AI can be applied in their fields.
The role of the ‘AI wrangler’
When he started at Google Labs, Johnson initially helped write prompts for AI models. As the AI boom took off, the role of prompt engineer similarly captured the attention of other English majors interested in AI.
Johnson said that prompting will change, especially as AI models improve at rewriting prompts — but a new role, which he referred to as the “AI wrangler,” has emerged.
“That’s maybe the next stage of the prompt engineer,” Johnson said.
Johnson describes the AI wrangler role as not necessarily requiring coding expertise but involving deep knowledge of the latest models and their capabilities.
For example, if someone wants to create a 30-second animated video with AI, the AI wrangler would know the best tool for that task and how to use it, Johnson said. He said the role requires a “certain level of technical sophistication” but doesn’t require knowing how to program.
Johnson said one of the most important skills to learn right now is fluency in the latest models and their functions.
“That’s just a general purpose skill that is actually going to be valuable in every single industry,” Johnson said.
Technical skills aren’t going away
While there may be a growing need for humanities skills to help build how models interact, that doesn’t mean that technical skills are no longer valuable or necessary. Those from a humanities background may need to skill up in technical areas to open up opportunities for themselves in the field.
A Google AI sales leader who studied journalism and started his career in filmmaking eventually joined the tech giant in its media, entertainment, and gaming sectors for AI sales. However, he told BI he had to earn technical certifications and work his way up in engineering roles before taking on AI leadership roles.
Johnson, a published author of 14 books, came to Google with a strong understanding of technology, having sold startups and written extensively on the subject.
“If you’re going to try and make API calls to a bunch of these models, you probably need some more technical skills,” Johnson said, adding that ” if someone wants to hire you to get the best outputs out of the best models, you don’t really need a lot of technical skills to do that.”