Question A:

Antitrust investigations of the dominant firms in artificial intelligence are likely to lead to substantially lower prices of AI products and services for businesses and consumers.

Responses weighted by each expert's confidence

Question B:

Antitrust investigations of the dominant firms in artificial intelligence are likely to promote greater competition and innovation in AI.

Responses weighted by each expert's confidence

Question C:

Potential harms from artificial intelligence are better assessed by market deployment rather than seeking to slow the pace of AI research and implementation.

Responses weighted by each expert's confidence

Question A Participant Responses

Participant University Vote Confidence Bio/Vote History
Acemoglu
Daron Acemoglu
MIT
Agree
5
Bio/Vote History
Although there is some uncertainty and government intervention could slow down innovation & harm the introduction of new services, the AI field right now is so oligopolistic that well-designed antitrust should be able to increase competition and reduce prices.
-see background information here
-see background information here
Aguiar
Mark Aguiar
Princeton
Uncertain
2
Bio/Vote History
Altonji
Joseph Altonji
Yale
Uncertain
3
Bio/Vote History
Auerbach
Alan Auerbach
Berkeley
Uncertain
3
Bio/Vote History
Autor
David Autor
MIT
Disagree
7
Bio/Vote History
The issue at stake is innovation, not price per se
Banerjee
Abhijit Banerjee
MIT
Uncertain
5
Bio/Vote History
Bergemann
Dirk Bergemann
Yale Did Not Answer Bio/Vote History
Bertrand
Marianne Bertrand
Chicago
Uncertain
1
Bio/Vote History
Brunnermeier
Markus Brunnermeier
Princeton
Uncertain
7
Bio/Vote History
Chevalier
Judith Chevalier
Yale
Uncertain
7
Bio/Vote History
Cutler
David Cutler
Harvard
Agree
4
Bio/Vote History
Duffie
Darrell Duffie
Stanford
Agree
1
Bio/Vote History
Edlin
Aaron Edlin
Berkeley
Uncertain
6
Bio/Vote History
Eichengreen
Barry Eichengreen
Berkeley
Uncertain
1
Bio/Vote History
Einav
Liran Einav
Stanford
Uncertain
1
Bio/Vote History
Fair
Ray Fair
Yale
Disagree
5
Bio/Vote History
Glaeser
Edward Glaeser
Harvard
Disagree
7
Bio/Vote History
Goldberg
Pinelopi Goldberg
Yale
Uncertain
6
Bio/Vote History
Greenstone
Michael Greenstone
University of Chicago
Uncertain
4
Bio/Vote History
Hart
Oliver Hart
Harvard
Uncertain
5
Bio/Vote History
Holmström
Bengt Holmström
MIT
Agree
4
Bio/Vote History
The key is to keep the open source community alive. I assume the antitrust investigation will advance that cause
Hoxby
Caroline Hoxby
Stanford
Agree
5
Bio/Vote History
Hoynes
Hilary Hoynes
Berkeley
Uncertain
5
Bio/Vote History
Hurst
Erik Hurst
Chicago Booth
Agree
4
Bio/Vote History
Judd
Kenneth Judd
Stanford
Uncertain
8
Bio/Vote History
What is the definition of an "AI product"? AI is jargon for a wide range of computational tools that extend standard numerical methods. Do we regulate statistical packages? If not, why regulate advanced data analysis tools?
Kaplan
Steven Kaplan
Chicago Booth
Disagree
8
Bio/Vote History
Kashyap
Anil Kashyap
Chicago Booth
Uncertain
3
Bio/Vote History
too many possible directions this could take to have any confidence on the outcome
Klenow
Pete Klenow
Stanford
Agree
2
Bio/Vote History
Levin
Jonathan Levin
Stanford
Uncertain
4
Bio/Vote History
Maskin
Eric Maskin
Harvard
Agree
5
Bio/Vote History
Nordhaus
William Nordhaus
Yale
Disagree
5
Bio/Vote History
Obstfeld
Maurice Obstfeld
Berkeley
Uncertain
1
Bio/Vote History
Pathak
Parag Pathak
MIT Did Not Answer Bio/Vote History
Samuelson
Larry Samuelson
Yale
Agree
8
Bio/Vote History
Scheinkman
José Scheinkman
Columbia University Did Not Answer Bio/Vote History
Schmalensee
Richard Schmalensee
MIT
Disagree
6
Bio/Vote History
No obvious case.
Scott Morton
Fiona Scott Morton
Yale
Disagree
5
Bio/Vote History
Shapiro
Carl Shapiro
Berkeley
Disagree
6
Bio/Vote History
Shimer
Robert Shimer
University of Chicago
Disagree
4
Bio/Vote History
This is a very new industry. Antitrust is more likely to stifle innovation and hence raise quality-adjusted prices
Stantcheva
Stefanie Stantcheva
Harvard
Uncertain
4
Bio/Vote History
Stock
James Stock
Harvard Did Not Answer Bio/Vote History
Stokey
Nancy Stokey
University of Chicago
Disagree
5
Bio/Vote History
It's a new industry, and there seems to be a lot of innovation and competition.
Syverson
Chad Syverson
Chicago Booth Did Not Answer Bio/Vote History
Thaler
Richard Thaler
Chicago Booth Did Not Answer Bio/Vote History
Udry
Christopher Udry
Northwestern Did Not Answer Bio/Vote History
Werning
Ivan Werning
MIT Did Not Answer Bio/Vote History

Question B Participant Responses

Participant University Vote Confidence Bio/Vote History
Acemoglu
Daron Acemoglu
MIT
Agree
5
Bio/Vote History
Once again there is some uncertainty, because successful antitrust would reduce the huge investments that OpenAI is planning. But overall, a more competitive field is likely to increase innovation and the quality of innovation. It may even enable a better direction of technology.
-see background information here
-see background information here
Aguiar
Mark Aguiar
Princeton
Uncertain
2
Bio/Vote History
Altonji
Joseph Altonji
Yale
Uncertain
3
Bio/Vote History
Auerbach
Alan Auerbach
Berkeley
Uncertain
3
Bio/Vote History
Autor
David Autor
MIT
Agree
7
Bio/Vote History
Banerjee
Abhijit Banerjee
MIT
Uncertain
5
Bio/Vote History
Bergemann
Dirk Bergemann
Yale Did Not Answer Bio/Vote History
Bertrand
Marianne Bertrand
Chicago
Agree
5
Bio/Vote History
Brunnermeier
Markus Brunnermeier
Princeton
Agree
6
Bio/Vote History
Chevalier
Judith Chevalier
Yale
Uncertain
6
Bio/Vote History
Cutler
David Cutler
Harvard
Agree
6
Bio/Vote History
Duffie
Darrell Duffie
Stanford
Uncertain
1
Bio/Vote History
Edlin
Aaron Edlin
Berkeley
Agree
6
Bio/Vote History
Eichengreen
Barry Eichengreen
Berkeley
Uncertain
1
Bio/Vote History
Einav
Liran Einav
Stanford
Uncertain
1
Bio/Vote History
Fair
Ray Fair
Yale
Disagree
5
Bio/Vote History
Glaeser
Edward Glaeser
Harvard
Strongly Disagree
8
Bio/Vote History
Goldberg
Pinelopi Goldberg
Yale
Uncertain
6
Bio/Vote History
There is fierce competition among the three biggerst AI pursuing firms - not clear antitrust is needed at this point
Greenstone
Michael Greenstone
University of Chicago
Uncertain
3
Bio/Vote History
Hart
Oliver Hart
Harvard
Uncertain
5
Bio/Vote History
Holmström
Bengt Holmström
MIT
Agree
5
Bio/Vote History
Hoxby
Caroline Hoxby
Stanford
Strongly Agree
8
Bio/Vote History
Hoynes
Hilary Hoynes
Berkeley
Uncertain
5
Bio/Vote History
Hurst
Erik Hurst
Chicago Booth
Agree
4
Bio/Vote History
Judd
Kenneth Judd
Stanford
Disagree
8
Bio/Vote History
One cannot affect the pace of research of computational tools.
Kaplan
Steven Kaplan
Chicago Booth
Disagree
7
Bio/Vote History
Dominant firms are intensely competitive and are investing intensely.
Kashyap
Anil Kashyap
Chicago Booth
Uncertain
3
Bio/Vote History
this seems more likely than not, but still too much ambiguity on the direction and what happens if activity and innovation are pushed outside the US?
Klenow
Pete Klenow
Stanford
Uncertain
1
Bio/Vote History
Levin
Jonathan Levin
Stanford
Uncertain
5
Bio/Vote History
Maskin
Eric Maskin
Harvard
Agree
4
Bio/Vote History
Nordhaus
William Nordhaus
Yale
Uncertain
5
Bio/Vote History
Obstfeld
Maurice Obstfeld
Berkeley
Agree
1
Bio/Vote History
Pathak
Parag Pathak
MIT Did Not Answer Bio/Vote History
Samuelson
Larry Samuelson
Yale
Agree
6
Bio/Vote History
Scheinkman
José Scheinkman
Columbia University Did Not Answer Bio/Vote History
Schmalensee
Richard Schmalensee
MIT
Disagree
6
Bio/Vote History
No obvious case.
Scott Morton
Fiona Scott Morton
Yale
Strongly Agree
8
Bio/Vote History
Shapiro
Carl Shapiro
Berkeley
Uncertain
8
Bio/Vote History
Shimer
Robert Shimer
University of Chicago
Disagree
5
Bio/Vote History
Stantcheva
Stefanie Stantcheva
Harvard
Uncertain
4
Bio/Vote History
Stock
James Stock
Harvard Did Not Answer Bio/Vote History
Stokey
Nancy Stokey
University of Chicago
Disagree
5
Bio/Vote History
Syverson
Chad Syverson
Chicago Booth Did Not Answer Bio/Vote History
Thaler
Richard Thaler
Chicago Booth Did Not Answer Bio/Vote History
Udry
Christopher Udry
Northwestern Did Not Answer Bio/Vote History
Werning
Ivan Werning
MIT Did Not Answer Bio/Vote History

Question C Participant Responses

Participant University Vote Confidence Bio/Vote History
Acemoglu
Daron Acemoglu
MIT
Disagree
5
Bio/Vote History
It is not either or. We need both more competition, and also much better regulation. Importantly, we need a redirection of AI in a more socially beneficial direction. To achieve this we may need to slow down the huge investments in data collection and moral training by incumbents
-see background information here
-see background information here
Aguiar
Mark Aguiar
Princeton
Uncertain
2
Bio/Vote History
Altonji
Joseph Altonji
Yale
Uncertain
3
Bio/Vote History
Auerbach
Alan Auerbach
Berkeley
Uncertain
3
Bio/Vote History
Autor
David Autor
MIT
Disagree
7
Bio/Vote History
Not every dangerous product should be first released on the market, then assessed for risks.
Banerjee
Abhijit Banerjee
MIT
Strongly Disagree
7
Bio/Vote History
Bergemann
Dirk Bergemann
Yale Did Not Answer Bio/Vote History
Bertrand
Marianne Bertrand
Chicago
Disagree
5
Bio/Vote History
Brunnermeier
Markus Brunnermeier
Princeton
Disagree
9
Bio/Vote History
Chevalier
Judith Chevalier
Yale
Uncertain
4
Bio/Vote History
Cutler
David Cutler
Harvard
Disagree
5
Bio/Vote History
Duffie
Darrell Duffie
Stanford
No Opinion
Bio/Vote History
Edlin
Aaron Edlin
Berkeley
Uncertain
7
Bio/Vote History
Eichengreen
Barry Eichengreen
Berkeley
Uncertain
1
Bio/Vote History
Einav
Liran Einav
Stanford
Strongly Agree
1
Bio/Vote History
Fair
Ray Fair
Yale
Agree
5
Bio/Vote History
Glaeser
Edward Glaeser
Harvard
Uncertain
3
Bio/Vote History
Goldberg
Pinelopi Goldberg
Yale
Agree
6
Bio/Vote History
Greenstone
Michael Greenstone
University of Chicago
Agree
2
Bio/Vote History
Hart
Oliver Hart
Harvard
Disagree
7
Bio/Vote History
There are many potential harms from AI and I don't think there is any reason to think the market will protect us from them--quite the opposite. We need a mixture of good corporate governance--often not in place-- and regulation.
Holmström
Bengt Holmström
MIT
Disagree
4
Bio/Vote History
But overbearing regulation should be avoided
Hoxby
Caroline Hoxby
Stanford
Uncertain
10
Bio/Vote History
Hoynes
Hilary Hoynes
Berkeley
Uncertain
5
Bio/Vote History
Hurst
Erik Hurst
Chicago Booth
Uncertain
3
Bio/Vote History
Judd
Kenneth Judd
Stanford
Uncertain
8
Bio/Vote History
The potential harms of AI are going to be difficult to counter by any means, just as it has been difficult to counter the bad effects of disinformation.
Kaplan
Steven Kaplan
Chicago Booth
Strongly Agree
8
Bio/Vote History
Information wants to be free. Experimentation will occur regardless of regulation. May as well let it happen.
Kashyap
Anil Kashyap
Chicago Booth
Uncertain
3
Bio/Vote History
Klenow
Pete Klenow
Stanford
Disagree
1
Bio/Vote History
Levin
Jonathan Levin
Stanford
Strongly Agree
5
Bio/Vote History
Maskin
Eric Maskin
Harvard
Uncertain
4
Bio/Vote History
Just as in the field of gene editing, there are probably some risks that can be assessed and safeguarded against before deployment
Nordhaus
William Nordhaus
Yale
Disagree
5
Bio/Vote History
Neither actually. Need oversight
Obstfeld
Maurice Obstfeld
Berkeley
Uncertain
1
Bio/Vote History
Pathak
Parag Pathak
MIT Did Not Answer Bio/Vote History
Samuelson
Larry Samuelson
Yale
Uncertain
1
Bio/Vote History
We have no experience assessing or controlling the potential harm of a technology like AI, and no reason to expect the market to be able to do so.
Scheinkman
José Scheinkman
Columbia University Did Not Answer Bio/Vote History
Schmalensee
Richard Schmalensee
MIT
Agree
5
Bio/Vote History
Scott Morton
Fiona Scott Morton
Yale
Uncertain
1
Bio/Vote History
Shapiro
Carl Shapiro
Berkeley
Agree
7
Bio/Vote History
Shimer
Robert Shimer
University of Chicago
Uncertain
5
Bio/Vote History
The main issue is that no country can unilaterally stop AI research and implementation. Slowing it down in the US increases the likelihood that researchers elsewhere develop the leading AIs, with potentially greater harm
Stantcheva
Stefanie Stantcheva
Harvard
Disagree
4
Bio/Vote History
Stock
James Stock
Harvard Did Not Answer Bio/Vote History
Stokey
Nancy Stokey
University of Chicago
No Opinion
Bio/Vote History
Syverson
Chad Syverson
Chicago Booth Did Not Answer Bio/Vote History
Thaler
Richard Thaler
Chicago Booth Did Not Answer Bio/Vote History
Udry
Christopher Udry
Northwestern Did Not Answer Bio/Vote History
Werning
Ivan Werning
MIT Did Not Answer Bio/Vote History