Print This Post
How’s this for an academic dream team? Albert Einstein, Niels Bohr and Richard Feynman. The three physicists and I often get together over coffee to chat about career decisions, difficulties and dilemmas. Of course, not the real Einstein, Bohr and Feynman — the last of the trio to die was Feynman, in 1988 — but rather, simulations of the three that I have assembled, using artificial intelligence (AI), to serve as my personal advisory board.AI is seemingly everywhere these days. It’s in productivity apps, web browsers and technical support lines. AI users increasingly report developing friendships, and even romantic relationships, with AI companions. I used the technology to develop an advisory board that can provide customized input into my career decisions and reflections. The resulting advice has been surprisingly useful — not to mention entertaining.
My board has provided advice on how to progress to my next career stage, how to contextualize my past and future work and how I might structure my next semester for maximum impact on my career. It also provided guidance when I received a job offer at a different university, helping me to see that accepting it might not be the best decision.
Although some people in corporate settings have proposed using AI to build a personal board of directors, I think the term ‘advisory board’ is more fitting, because the AI panellists are only advisers — they should not make decisions for you, and they cannot own the consequences of their decisions.
The make-up of my advisory board often changes, and has included an eclectic mix. Besides Bohr, Feynman and Einstein, I’ve also tapped microbiologist Alexander Fleming, poet Piet Hein and anti-apartheid activist Nelson Mandela. Sometimes I include experts from my specific disciplines of AI and marketing; other times, I ‘invite’ artist Pablo Picasso or architect Bjarke Ingels for a completely different perspective. But whatever the board’s composition, I typically retain a core group of three seminal scientists.
The make-up of the board influences not just what I might be told, but also the mood in the room. Picasso might push for an intuitive representation or metaphor, whereas Bohr tends to insist on complementarity and paradoxes. Different ‘personalities’ might dominate depending on the question — but, unlike with human discussions, you can often prompt your way out of any issues. You can, for example, prompt the AI to synthesize the board’s answers into a single coherent piece of advice, or to fuel ‘constructive conflict’ between members. As a result, I sometimes find myself acting more as a moderator than a passive recipient of information.
Build-a-board workshop
So, how do you build your own AI-generated board?
First, choose your favourite chatbot, such as ChatGPT (provided by OpenAI), Claude (Anthropic) or LeChat (Mistral AI). Choosing one that has some memory of your previous conversations can be helpful. My profile remembers that I’m an associate professor, as well as the specific challenges I face and my interests. As a result, I don’t have to start every conversation from scratch.
Next, prompt the AI to act like an academic advisory board. Here’s a prompt I might use to initiate an AI brainstorming session:
“Assemble an imaginary advisory board consisting of [INSERT PEOPLE HERE]. Have them discuss my current academic situation and give me contrasting, thought-provoking advice. End with a short synthesis that integrates their perspectives into one actionable insight.”
Of course, this is only a starting point. The real value lies in continuing the conversation.
You could also try presenting the board with a dilemma you’re facing and asking for input.
Finally, experiment and tinker with the make-up of your board. Try including different personalities, such as talk-show host Oprah Winfrey or author Stephen King. The sheer diversity can often reveal surprising perspectives on the challenges you face. For instance, in response to the generic prompt above, ‘Nelson Mandela’ reminded me that institutions often follow people of integrity, so I should focus on lifting other people up and ask whose lives will be improved if I succeed in my academic career.
Danger on board
There are, of course, pitfalls to consider when creating your own advisory board.
First, never outsource your decision-making. Use the technology to ‘open up the problem’ to perspectives that differ from your own, and, through this, to aid your own decision-making on a career-related matter. Because an AI can never take responsibility for its decisions (and might not have all the relevant information), you must always make the final decision yourself.
Second, trust your instincts. AI can help you to see new perspectives and consider variables you might not have thought of, allowing you to make a rational and informed decision. But whatever decision you make, it should also feel right. Don’t let the AI drown out your inner voice.
Third, include an eclectic mix of people. My rule of thumb is that prospective panellists should have something to say (a valuable perspective) and something to add (a unique perspective). Mismatches can produce amusing, although not always helpful, discussions. I once tried to include Alexander Fleming (who discovered penicillin) on my advisory board, but accidentally added James Bond creator Ian Fleming instead. ‘Ian Fleming’ suggested that I was more than a professor: I was a secret agent with a cover (a polished professor with creative ideas beneath the surface), cause (an underlying mission or driving purpose) and code (a set of personal rules, such as, ‘never publish what you don’t believe’).
Fourth, remember that your board members are digital stereotypes. They might represent only one dimension of the people on whom they are based, not those individuals’ real personalities. I know this because I’ve also tried adding famous colleagues whom I know personally. Their AI counterparts provided perfect input according to some of their published works, but didn’t answer questions as they would have done. Indeed, the real-world researchers might not even agree with the advice their AI counterpart gives on an issue.
Finally, don’t use the AI-enabled advisory board as a shortcut to finding answers, but rather to ask better questions. Part of the challenge when it comes to making career decisions is that the future is unknown. An AI can’t remove this uncertainty, but it can help you to explore different options. However, it might also confidently provide you with answers as if it knows what will happen, so treat it as an instrument for exploration rather than a crystal ball.
Generative AI is a powerful technology with both promise and problems. It can be either intelligent or ignorant, depending on how we use it. Put differently, maybe we need a form of meta-intelligence to guide us on how to use our own and machine intelligence intelligently.
And in case you were wondering, yes, ‘Einstein’, ‘Bohr’ and ‘Feynman’ all thought it was a good idea for me to write this piece.
doi: https://doi.org/10.1038/d41586-025-03468-9
This is an article from the Nature Careers Community, a place for Nature readers to share their professional experiences and advice. Guest posts are encouraged.
Subscribe to Nature Briefing: Careers, an unmissable free weekly round-up of help and advice for working scientists.
Nature Careers Blog
