Discover what AI can’t do — even as it dazzles — in this insider look at Columbia’s inaugural AI Summit.
I’ll never forget walking into Low Library on March 4, 2025 — bright-eyed, coffee in hand — ready to drink from the firehose of AI enthusiasm. Instead, I was struck by a paradox: the more capable these models become, the clearer their limits. Scholars from architecture to sociology gathered to ask, essentially, “Can a machine really understand us…or even care?” Let’s unpack seven tensions that emerged when the world’s brightest minds convened at Columbia’s first AI Summit.
By convening thousands of researchers, technologists, and students across the Morningside, Manhattanville, and medical campuses, Columbia signaled that AI isn’t siloed — it’s everywhere, reshaping medicine, energy, design, law, and the arts. Organized by the Data Science Institute under the Columbia AI initiative (shout‑out to Garud Iyengar!), this was not a rah‑rah tech fest but a sober reflection on responsibility.