Anjana Ahuja’s thoughtful column “Why the right questions in science can be the hardest” (Opinion, November 19) highlights a truth that is becoming increasingly urgent in the age of artificial intelligence. As AI improves at generating answers — sometimes before we even finish typing the question — the real bottleneck is shifting. What societies now lack is not answer-literacy but question-literacy.
Over the past decade, we have invested heavily in data science, computational methods, and STEM subject education (science, technology, engineering and mathematics). Yet we have devoted remarkably little attention to cultivating the capability that precedes all discovery: the ability to formulate meaningful, ambitious, and socially relevant questions.
In my own work on developing a new science of questions, I have argued that asking better questions is not simply a precursor to knowledge creation; it is a civic and democratic competence.
For that reason, while Harvard’s “Genuinely Hard Problems” initiative is commendable, the article misses a deeper structural issue: who gets to decide which questions are worth asking. If only elite institutions curate the scientific agenda, we reproduce what I call “question inequities”: the tendency for a small group of researchers, often from the same disciplines, countries, or demographics, to determine what counts as a “hard problem”. In an era facing polycrises, this narrowing of curiosity is a luxury we cannot afford.
We need a more participatory model for question formation, involving not just scientists but policymakers, practitioners and communities whose lived experience often reveals gaps invisible to academia. That principle guided The GovLab’s “100 Questions Initiative”, which engaged global networks to identify the most pressing questions across domains such as mobility, migration, gender and governance. A similar effort — a “100 Questions for Society” — could complement the Harvard initiative by ensuring that the questions shaping science are not only intellectually challenging but socially consequential.
If AI is increasingly answering questions, our true competitive advantage will lie in asking the right ones — and ensuring everyone has a voice in determining what they are.
Stefaan Verhulst
Research Professor, Tandon School of Engineering, New York University, Co-Founder, The GovLab (NYC) and The DataTank (Brussels), New York, NY, US