Do you think we no longer need PhD students if AI can do PhD-level research?
Before I answer that, let’s go to what Sam Altman told his little brother, Jack Altman on his podcast (uncapped) last week:
“Whatever happens with AI in the future, we’ll always need humans in the story. Only humans can say, ‘I made this decision. I made that mistake.’ Because humans care about humans.”
Now one more example from my role before I go back to the PhD student topic. What I do is a high-tech role in a top tech company. Since I was granted access to internal AI assistants, I have been able to do 10x more. I created AI to agents do what I used to do and now manage automated workflows at an aggregated level. So 10x more productivity, and the company sees no reason to fire me but instead, they need more people like me who know how to use AI to do more.
Now, back to PhD students. AI can outperform PhD students in advanced research tasks and in novel ideas meaning a PhD student who can leverage AI can operate at a professor level, orchestrating multiple “AI-PhDs” and managing a portfolio of projects. A professor who had 10 PhD students previously can now be a super-professor managing 10 professors with at least 10x more PhD-level research volume, and with quality increasing over time (since AI’s capabilities improve as they learn and models advance).
This is a huge win for humanity: higher-quality research projects, more research projects, and solutions to previously unsolved problems.
This does not mean we no longer need PhD students. It’s the opposite. We need more PhD students who can orchestrate and leverage AI, who will be the humans making final decisions, publishing outcomes, and claiming credit for the work.
Having AI doing “our job” means we now have a superhero tools that can do our job even better, but it still needs us to orchestrate those super-tools and benefit from them in our area of expertise. And then, we can do heroic things that were not possible before.
Mohsin With E 
Leave a comment