How do people feel about AI?

Website documentation

Chart content blocks

Figure 2: Self-reported personal experience with each of the AI uses ordered from most to least experience

People report the highest experience for targeted consumer ads and smart speakers, and the lowest for cancer risk detection and calculating welfare eligibility. Percentages may not add to 100 due to rounding.

​“How much personal experience have you had with…” 

Targeted consumer advertising
A lot: 37%
Targeted consumer advertising
Some: 44%
Targeted consumer advertising
Not sure/​Prefer not to say: 4%
Targeted consumer advertising
None: 16%
Smart speakers
A lot: 24%
Smart speakers
Some: 40%
Smart speakers
Not sure/​Prefer not to say: 4%
Smart speakers
None: 32%
Facial recognition for unlocking phones
A lot: 34%
Facial recognition for unlocking phones
Some: 28%
Facial recognition for unlocking phones
Not sure/​Prefer not to say: 1%
Facial recognition for unlocking phones
None: 37%
Facial recognition for border control
A lot: 18%
Facial recognition for border control
Some: 41%
Facial recognition for border control
Not sure/​Prefer not to say: 2%
Facial recognition for border control
None: 39%
Targeted political advertising
A lot: 12%
Targeted political advertising
Some: 36%
Targeted political advertising
Not sure/​Prefer not to say: 10%
Targeted political advertising
None: 42%
Virtual healthcare assistants
A lot: 4%
Virtual healthcare assistants
Some: 24%
Virtual healthcare assistants
Not sure/​Prefer not to say: 8%
Virtual healthcare assistants
None: 65%
Assessing loan repayment risk
A lot: 3%
Assessing loan repayment risk
Some: 21%
Assessing loan repayment risk
Not sure/​Prefer not to say: 9%
Assessing loan repayment risk
None: 66%
Virtual reality in education
A lot: 3%
Virtual reality in education
Some: 19%
Virtual reality in education
Not sure/​Prefer not to say: 6%
Virtual reality in education
None: 71%
Robotic vacuum cleaners
A lot: 6%
Robotic vacuum cleaners
Some: 15%
Robotic vacuum cleaners
Not sure/​Prefer not to say: 1%
Robotic vacuum cleaners
None: 78%
Assessing job eligibility
A lot: 3%
Assessing job eligibility
Some: 14%
Assessing job eligibility
Not sure/​Prefer not to say: 7%
Assessing job eligibility
None: 76%
Facial recognition for policing
A lot: 2%
Facial recognition for policing
Some: 10%
Facial recognition for policing
Not sure/​Prefer not to say: 5%
Facial recognition for policing
None: 83%
Assessing welfare eligibility
A lot: 2%
Assessing welfare eligibility
Some: 9%
Assessing welfare eligibility
Not sure/​Prefer not to say: 7%
Assessing welfare eligibility
None: 82%
Assessing risk of cancer
A lot: 2%
Assessing risk of cancer
Some: 6%
Assessing risk of cancer
Not sure/​Prefer not to say: 5%
Assessing risk of cancer
None: 88%
Download CSV
Technology A lot Some Not sure/​Prefer not to say None
Targeted consumer advertising 36.74220343% 43.91380237% 3.832197267% 15.51179693%
Smart speakers 24.43264716% 39.66710993% 3.696975615% 32.2032673%
Facial recognition for unlocking phones 33.52157811% 28.36484648% 0.656678667% 37.45689674%
Facial recognition for border control 18.32073198% 40.60572152% 1.948501458% 39.12504504%
Targeted political advertising 12.38176208% 35.71367945% 9.682222078% 42.22233639%
Virtual healthcare assistants 3.562494819% 24.01657098% 7.633910075% 64.78702412%
Assessing loan repayment risk 3.449049271% 21.43966235% 9.428198036% 65.68309035%
Virtual reality in education 3.477922355% 19.29012413% 6.310642276% 70.92131124%
Robotic vacuum cleaners 5.573817071% 14.65572794% 1.316677426% 78.45377756%
Assessing job eligibility 2.617093394% 13.87717676% 7.119662635% 76.38606721%
Facial recognition for policing 2.384167031% 9.890629892% 5.040263707% 82.68493937%
Assessing welfare eligibility 2.002663071% 8.777482118% 7.101915985% 82.11793883%
Assessing risk of cancer 1.646809372% 6.228682127% 4.561317657% 87.56319084%

Table chart

Chart description lorem ipsum dolor sit amet, consectetur adipiscing elit. Id enim sapien massa fusce.

The question asked in the survey. 

Technology
Top three chosen benefits
1st
2nd
3rd
Assessing risk of cancer
1st
Earlier detection
82%
2nd
Less human error
53%
3rd
More accurate
42%
Assessing loan repayment risk
1st
Faster and easier
52%
2nd
Less likely to discriminate
39%
3rd
Less human error
37%
Technology Top three chosen benefits Percentage
Assessing risk of cancer 1 Earlier detection 82%
2 Less human error 53%
3 More accurate 42%
Assessing loan repayment risk 1 Faster and easier 52%
2 Less likely to discriminate 39%
3 Less human error 37%

Bar chart

Chart description lorem ipsum dolor sit amet, consectetur adipiscing elit. Id enim sapien massa fusce.

The question asked in the survey. 

Laws and regulations
Net benefit: 62%
Appeals process for AI decisions
Net benefit: 59%
Security of personal information
Net benefit: 56%
Explaining AI decision-making
Net benefit: 54%
Equal treatment monitoring
Net benefit: 53%
More human involvement
Net benefit: 44%
Government regulator approval
Net benefit: 38%
Don’t know/​Prefer not to say
Net benefit: 3%
Nothing
Net benefit: 3%
Something else
Net benefit: 1%
None of these
Net benefit: 1%
Technology Net benefit
Laws and regulations 62%
Appeals process for AI decisions 59%
Security of personal information 56%
Explaining AI decision-making 54%
Equal treatment monitoring 53%
More human involvement 44%
Government regulator approval 38%
Don’t know/​Prefer not to say 3%
Nothing 3%
Something else 1%
None of these 1%

Figure 5: Net benefit scores for each technology

A positive score indicates benefits outweigh concerns, while a negative score indicates concerns outweigh benefits.

Independent regulator
1.5
Something else companies
0.6
Autonomous weapons
-0.52
Technology
Independent regulator 1.5
Something else companies 0.6
Autonomous weapons -0.52

Figure 10: Who the public thinks should be most responsible for ensuring AI is used safely, split by age

​“Question asked in survey” 

60%
40%
20%
0%
18–24
25–34
35–44
45–54
55–64
65–74
75+ 
Other
18–24 years: 1%
Other
25–34 years: 2%
Other
35–44 years: 6%
Other
45–54 years: 8%
Other
55–64 years: 3%
Other
65–74 years: 5%
Other
75+ years: 7%
Other
18–24 years: 0%
Other
25–34 years: 1%
Other
35–44 years: 2%
Other
45–54 years: 4%
Other
55–64 years: 1%
Other
65–74 years: 0%
Other
75+ years: 8%
AI technology companies
18–24 years: 43%
AI technology companies
25–34 years: 36%
AI technology companies
35–44 years: 29%
AI technology companies
45–54 years: 24%
AI technology companies
55–64 years: 18%
AI technology companies
65–74 years: 15%
AI technology companies
75+ years: 19%
Independent regulator
18–24 years: 25%
Independent regulator
25–34 years: 34%
Independent regulator
35–44 years: 36%
Independent regulator
45–54 years: 46%
Independent regulator
55–64 years: 49%
Independent regulator
65–74 years: 52%
Independent regulator
75+ years: 38%
Age group 18–24 25–34 35–44 45–54 55–64 65–74 75+
Independent regulator 25% 34% 36% 46% 49% 52% 38%
AI technology companies 43% 36% 29% 24% 18% 15% 19%
Other 1% 2% 6% 8% 3% 5% 7%
0% 1% 2% 4% 1% 0% 8%