MGAIC Kickoff symposium

Wow, what an amazing turnout! Thank you, Anantha, for the introduction – and for your outstanding leadership in bringing the consortium to life. 

• A warm shout-out to our faculty co-directors, Vivek Farias – and Tim Kraska – and to the whole team of faculty and staff who organized today’s program.
• Thanks to all the students and early-career researchers who’ll share their work in this afternoon’s poster session.
• And I know we’re all looking forward to hearing from today’s incredible speakers and panelists, starting with the legendary Yann LeCun.

Today’s symposium is the first of what will be an annual flagship event for the MIT Generative AI Consortium. Now, at MIT, the one thing we love as much as we love numbers is a good acronym. UROP! MISTI! DAPER! DUSP! As you may have noticed, in this case, with a little acronymical sleight of hand, we decided to pronounce M-G-A-I-C as “magic.”

Easy to remember and fun to say!  Can’t argue with that.

But I would argue that calling this effort “magic” also points to something quite serious, about why we’re all here together.

Now, in this audience, there’ve got to be some fans of the great sci-fi master, Arthur C. Clarke, yes?

Remember Clarke's Third Law?

“Any sufficiently advanced technology is indistinguishable from magic.” 

I know I’m not the first to connect Arthur Clark’s observation with the “black box” sensation of using Generative AI: Put in your prompt, and you get a polished answer back so fast that it surely feels like magic.

Of course, everyone here knows that Gen AI doesn’t work by magic. 

And yet – even the experts can’t always explain how it does work. That makes GenAI fundamentally different from essentially any other technology. And it creates a very important responsibility for an institution and a community like MIT.

This audience includes experts, and aspiring experts, on every part of the AI menu. And I am certainly not one! But to explain the kind of responsibility I’m talking about, let me offer, as an example, my own field: the life sciences. 

My lab explored the deep behavior of cancer cells. Today, when I talk to cancer researchers and clinician-scientists on our faculty, they say that AI already touches – and improves – literally every aspect of their work, from how they design experiments and analyze data to how they keep and manage records to how they diagnosis and care for their patients. 

They’re positioned to capitalize on a great many AI advances – right now. And I hear tremendous gratitude and enthusiasm. So part of MIT’s responsibility is to keep those advances coming! 

But those researchers and clinicians, on their own, are not in a position to address AI’s risks and mysteries. How to guarantee accuracy and guard against hallucinations? How to get sufficient patient data to make the systems work, while preserving confidentiality? How to validate the results and achieve enough “explicability” that users feel safe to trust them? 

In short – how can you manage the “magic” so that all of us can confidently rely on it for critical applications in the real world?

And so, they’re counting on all of you – our faculty, researchers and students, together with our consortium members – for help. They’re counting on you to tackle the whole suite of technical and ethical challenges whose solutions will allow society to make the most of this astounding technology. And – in the myriad other realms where AI is also changing everything – the rest of society is counting on you too.

It's a big responsibility. But, judging by the depth and breadth of interest on campus, I have every confidence that you’re up to it! 

The excitement is everywhere: The two new undergraduate majors we’ve created in these fields were effectively “sold out” overnight. And when we announced our first round of MGAIC seed funding, we were stunned to receive more than 180 research proposals, representing every School and the College. They cover an incredible range of ideas – from fundamental research to practical applications, from responsible deployment to workforce transformation. Many depend on collaborations across departments and disciplines – and, crucially, collaborations with industry, too.

So I want to close with a special nod to the consortium’s founding member companies – Analog Devices, Coca-Cola, InterSystems, OpenAI, SK Telecom, Tata and TWG Global. What enables the “magic” in MGAIC is your commitment to explore very important AI challenges with us – in the best “open source” tradition of MIT. We wouldn’t be here today without you – and we’re deeply grateful for your willingness to join us in this vital work.

Thank you all for coming! And best wishes for a fascinating day!