HAI Roundtable on AI and the Economy

…the first event at Stanford’s HAI (Human-Centered AI Institute) today.

We were asked to open with the prompt: What do you see as the top three challenges and opportunities that AI presents for the Economy? I suggested:

1) The inevitability of accelerating inequality in our current economic system
2) Concentration effects in companies (Google ASIC farms) and cultures (China, unfettered by privacy laws like HIPAA), compounding power laws on fractal scales.
3) The abundant opportunity in a post-employment era
4) Evolutionary divergence in technology (between classical engineering and iterative algorithms)

“If we are always learning in our work, it is one form of meaning. Can we be constantly learning and constantly skilling versus re-skilling.” — John Seely Brown

“Somewhere along the way, policy got divorced from evidence, and then got married to identity.” — Scott Phoenix, Vicarious

“We have a crisis of confidence in technology” — Mustafa Suleyman, DeepMind

“AI is often trained with the goal of meeting or exceeding targets of human ability. So, they end up being substitutes for human activity, rather than complements. Could we build very different machines that have goals to complement us and not do what we already do?” — James Manyika, McKinsey

“AI is born of original sin. We are a very technical field where representation is massive skewed. My singular issue is diversity, inclusion, and social equity in AI. This is linked to the fate of AI.” — Fei-Fei Li, Stanford HAI

“I like democracy, but it can’t pivot.” — Andrew Ng

“Education is different from learning. Learning is fundamentally social.
Harvard did a study of their graduates. The best indicator of future career success was the willingness to form or join a study group.” — JSB

“In the past 8 years I have received zero NSF funding. Most of it comes from companies. Stanford is one of less than 10 universities that can enjoy industry funding. I am freaking out about this. The majority of research institutions don’t have nearly the same level.” — Fei-Fei Li, Stanford HAI

I’ll share the summary policy/research agenda paper when it becomes available. Meanwhile, here is the summary from the last one, 4 years ago (that I hosted at work).

With: Erik Brynjolfsson (MIT Initiative on the Digital Economy), Reid Garrett Hoffman, Fei-Fei Li (HAI) and James Manyika (McKinsey Global Institute), in this lineup, and John Seely Brown, Andrew Ng, Eric Schmidt, Myron Scholes , and Mustafa Suleyman (Deep Mind) among others.

Fei-Fei launched the HAI Institute earlier this week… with a Rethink Robot at hand (WIRED)