Polis 2.0

Understanding Americans’ Views on AI

​

​

Maximilian Kroner Dale

Darshana Narayanan

Colin Megill

Recruited over 1,000 quota-sampled Americans

• Representative on
• Gender (50% female),
• race & ethnicity (63% white including hispanic, 47% non-white),
• Region (40% south, 22% west, 21% midwest, 17% northeast)
• Urbanicity (21% rural, 48% suburban, 31% urban)
• Age (except too few 18-24 yos)
• Income
• Deviations:
• Slight over-representation of Dems
• Slight under-rep of 18-24 yo

Voters by time of first vote

87,317 votes cast, 77 votes per person on average

• ~150 seed comments�
• ~1000 user-written statements, 300 other comments removed by human moderation�
• Almost 90,000 total votes

Topics identified as most important by participants

1. Privacy & Surveillance

2. Regulation & Accountability

"AI-powered surveillance is a huge threat to personal privacy and basic rights."

​

"There needs to be protections from AI monitoring and consent from people. Especially protections so their likeness can't be kept or used."

"AI systems must be transparent and accountable to prevent misuse and ensure public trust."

​

"Legislation needs to be updated on several fronts. Trademark, proliferation, content regulation for deep fakes."

3. Deepfakes & Misinformation

4. Job Displacement

"AI could seriously mess with democracy by spreading fake news and targeted manipulation."

"My biggest concern about Artificial Intelligence is the effect of employment it will bring to people. People will lose their jobs due to AI."

​

5. Education & Critical Thinking

"Kids should learn basic skills and critical thinking before being taught how to use AI."

Polis clustered participants into two Groups

Group A

Group B

Statements that illustrate this divide:

• Larger: 773 people
• Includes higher proportion of Democrats and Independents
• Very interested in regulation all kinds for AI
• Concerned about AI’s current impact on our humanity

• Smaller: 273 people
• More often are Republican (+12.7% more)�
• A bit more concerned about gov’t overreach
• More optimistic about the potential for AI

Overall

Group A

Group B

Plenty of bridging statements: �Many views were shared by both groups A & B

1. Telling Truth from Fiction

2. Deepfakes & Data Privacy

3. Kids Education & Critical Thinking

4. Need for Accountability & Oversight

“Social media should clearly mark AI-made posts so people know what’s real.”

“AI can be used as a supplemental learning material. But kids need to be taught how to identify and think about things before using it.”

​

“Deepfakes are making it harder to believe what we see online.”

“AI must be regulated to prevent misuse, protect privacy, and ensure ethical development over the next decade.”

Standard Report

Topic Visualization

Topic Statistics

Consensus Statement:

Topic 2 - AI’s Impact on Employment

Consensus Statement Snippet

We agree it is the future of technological advances¹⁸² though we understand it will cause some people to lose jobs while potentially benefiting jobs in other industries¹⁵⁴

^�We see AI as a cost-cutting measure utilized by large corporations⁶³² and believe that if a company replaces people with AI, they should at least help those workers retrain or give fair severance⁷⁷

​

We recognize that office workers who don't learn AI tools are going to fall behind⁸¹ and that education systems need to prepare students to work alongside AI, not compete with it⁸⁸⁶

Come find us if you are interested in the reports, �or curious about Polis 2.0

Maximilian Kroner Dale

​

Darshana Narayanan

​

Colin Megill

max.kronerdale@gmail.com

​

darshana@compdemocracy.org

​

colin@pol.is

​

Conflict & Existential Risk