MIT should strategically pause undergraduate education to focus on AI
MIT should strategically pause undergraduate education for two years to focus on AI.
As the greatest university in the world – and far and away the best science and engineering university – only MIT has the skill and credibility to lead the world as it navigates the technological, political, and economic changes that AI will bring.
AI is both the greatest opportunity and the greatest danger of our time.
Ray Kurzweil ’70 believes that intelligence will expand by a factor of a million within twenty years. He says that we will merge with AI and live forever (barring accidents). And that our consciousness will expand.
Others believe that AI will escape human control and turn humans into pets. Even before achieving superhuman intelligence, AI could lead to mass inequality and unemployment. Or, as Ilya Sutskever has suggested, it could create “infinitely stable dictatorships.”
As the greatest collection of scientists and engineers ever assembled, MIT is uniquely positioned to both develop and apply AI across the greatest range of disciplines and to harness and yoke the technology to reduce the probability of catastrophic harm.
Unfortunately, the faculty of MIT is hampered by its undergraduate responsibilities. Teaching courses and mentoring undergraduates, while noble, is a long-term investment. The opportunities and challenges presented by AI are here and now and humanity will not get a second chance if it fails the first time.
I will give an example of MIT failing to rise to the occasion to illustrate the importance of acting now. I spent most of the summer of 2023 trying to convince MIT to take a position on fake AI images and videos heading into the 2024 presidential election. I had three face to face meetings in which I made the case that AI presented a clear and present danger to the 2024 elections and that MIT was uniquely positioned to serve the country. Unfortunately, there was no follow up - I failed. And now there are fake videos about candidates that have over 100M views. I feel confident that the situation will get even worse as we approach the election.
So, to give MIT faculty, scientists, and engineers the greatest amount of time to focus on AI, I recommend that MIT strategically pause undergraduate education for two years. This will allow faculty to focus on AI without distractions.
Some may argue that undergraduate education is key to MIT’s mission. That is true but to preserve that mission it must be temporarily paused. AI threatens all of humanity so a two year pause is a small price to pay to reduce the AI threat.
Others will argue that MIT needs the tuition from undergraduates to fund its operations. That is true in normal times, but now MIT should dip into its $20B+ endowment to fund AI research and policy recommendations.
While my proposal to pause undergraduate education for two years is radical, it is necessary. But if MIT does not have the courage to adopt it, I suggest two smaller, but impactful ideas.
First, MIT should start an AI Fellows program to keep AI talent at MIT. Every year hundreds of AI-smart graduate students, postdocs, and research scientists leave MIT when they get their degrees or their appointments end. The AI Fellows program would keep these scientists and engineers at MIT and fund applied work in all areas from archaeology to zoology. Each AI Fellow would choose the field of greatest interest to them and then apply AI to accelerating science in that field.
Second, MIT should take AI as seriously as it is taking climate change. Compared to AI, climate change is irrelevant: if AI turns out to not be aligned with human interests, it will end humanity far before climate change does. Metaculus predicts that artificial general intelligence will be achieved by 2032, far before the worst effects of climate change will be felt.
And if AI is aligned with human interests it will help us solve climate change[2][3]. So MIT should create an AI Institute modeled on its climate change initiative.
Not since World War II has the world needed MIT as badly as it does now. It is time to act.
Thanks to Kriss Aho ‘83 for commenting on a draft.
Michael de la Maza SB’92, SM’93, PhD’97 is best known for beating Sam Altman in a 5 year, $100K bet on startup valuations. He lives in Cambridge, MA.