People want AI to fill gaps, not replace roles
If people don't want AI taking jobs, where do they want it? The answer is strikingly consistent. When asked what would make AI beneficial for their community, respondents converge on the same three themes: creating and protecting jobs, making healthcare accessible, and improving education quality (Chart 16). This trio appears in the top three for seven of the ten regions surveyed, with the order varying but the priorities remaining stable. These priorities also track closely with what people say matters most for their own wellbeing: healthcare (43%), food and water (38%), and meaningful work (35%; Chart 8).
What would AI have to do to be considered beneficial?
Coded themes from open-ended responses on what would make AI beneficial (n=1,030)
Create/protect jobs
21%
Make healthcare accessible
21%
Improve education quality
21%
Be accessible to all
16%
Be affordable/free
12%
Be controlled locally
10%
Improve government/public services
8%
Fight corruption
7%
Provide clean energy/water
6%
Ensure safety/privacy
6%
Reduce inequality
5%
Improve infrastructure
5%
Help with farming
3%
Work in local languages
1%
0
5
10
15
20
25
Struggling respondents report much lower satisfaction across every domain than comfortable ones, with gaps as wide as 0.87 points on meaningful work and 0.82 on housing (Chart 5). But their expectations of AI narrow that gap considerably (Chart 6). On healthcare, for example, struggling respondents rate current satisfaction at 2.80 but expect AI impact at 3.60, a jump of 0.80 points. Comfortable respondents start higher (3.47) but expect a smaller uplift (to 3.97). The pattern holds across most domains. Those with the least are not the most cynical about AI; they are less optimistic in absolute terms, but they see more room for improvement and expect AI to deliver some of it. The exception is community, where struggling respondents rate AI's expected impact (2.69) below their current satisfaction (2.78), the only domain for any income group where people expect AI to make things actively worse.
How needs are met across income levels?
Current satisfaction by household income situation — mean scores on 5-point scale
Healthcare
Education
Food & water
Housing &
infrastructure
Safety
Governance
Social
support
Meaningful
work
Environment
Community
Leisure time
Struggling
45
2.8
3.0
2.9
2.4
2.6
2.1
2.4
2.2
2.4
2.8
2.6
Stretched
226
3.1
3.2
3.2
2.9
2.9
2.4
2.6
2.5
2.6
2.8
3.0
Getting by
474
3.3
3.4
3.3
3.1
3.0
2.6
2.9
2.8
2.8
2.9
3.2
Comfortable
296
3.5
3.5
3.4
3.2
3.1
2.8
3.1
3.1
3.0
3.2
3.4
Not met
Fully met
Respondents are most optimistic about AI’s impact on healthcare, education, and leisure time, and most pessimistic about its effects on the environment, community bonds, governance, and employment (Chart 6). The top and bottom of this ranking are revealing. The domains where people expect AI to help most are those that can be improved through better information and efficiency: diagnosing illness, personalising learning, automating routine tasks. The domains where expectations are lowest involve human relationships, cooperation, collective decision-making and a sense of purpose, things that are harder to scale or automate.
How do people expect AI to change things?
Expected AI impact (B4) by household economic situation — mean scores (1=much worse, 3=same, 5=much better)
Healthcare
Education
Food & water
Housing &
infrastructure
Safety
Governance
Social
support
Meaningful
work
Environment
Community
Leisure time
Struggling
45
3.6
3.8
3.1
3.0
3.2
2.6
3.1
2.8
2.6
2.7
3.2
Stretched
226
3.8
3.9
3.2
3.4
3.3
2.9
3.2
3.0
2.9
2.7
3.6
Getting by
474
3.9
3.9
3.3
3.4
3.3
3.0
3.2
3.0
3.0
2.8
3.6
Comfortable
296
4.0
4.0
3.4
3.5
3.4
3.1
3.4
3.2
3.1
3.1
3.7
Not met
Fully met
The hope matrix: where can AI help most?
Current satisfaction (B3) vs. expected AI impact (B4) — overall (n=1,068)
MATRIX CHART PENDING
Survey participants were asked: If AI automates many jobs and people spend less time working, which would you most want for your community? 57% choose free public services, far ahead of lower prices (23%), direct cash (10%), or AI tools for communities (11%; Chart 10). The preference for public services holds across every income group and every region in the survey, with remarkably little variation across income brackets (53-59%). This is a notable finding given ongoing policy debates around universal basic income. When offered the choice, people across very different economic circumstances prefer collective provision over individual transfers.
If AI automates jobs, what do you want the most?
Percentage selecting each option as top preference — by household economic situation
Free public services
57%
Lower prices on goods
23%
Cash distributed directly
10%
AI tools for communities
11%
0
10
20
30
40
50
60
70
Free public services
Overall
1,041
57%
Struggling
45
53%
Stretched
226
54%
Getting by
474
57%
Comfortable
296
59%
0
10
20
30
40
50
60
70
Lower prices on goods
Overall
1,041
23%
Struggling
45
24%
Stretched
226
27%
Getting by
474
22%
Comfortable
296
20%
0
10
20
30
40
50
60
70
Cash distributed directly
Overall
1,041
10%
Struggling
45
11%
Stretched
226
12%
Getting by
474
9%
Comfortable
296
9%
0
10
20
30
40
50
60
70
AI tools for communities
Overall
1,041
11%
Struggling
45
11%
Stretched
226
7%
Getting by
474
11%
Comfortable
296
13%
0
10
20
30
40
50
60
70
The regional picture is broadly similar, though direct cash payments find more support in Eastern Europe (22%) and North America (21%) than in Southeast Asia (2%) or Sub-Saharan Africa (3%), where public service provision may feel like a more pressing gap to fill. Sub-Saharan African respondents are the most enthusiastic about AI tools for communities (19%, tied with their preference for lower prices), suggesting that where public service infrastructure is weakest, people see more value in direct tool provision rather than expanded government programs.
That Sub-Saharan Africa and Southeast Asia are the least interested in cash transfers is perhaps the most counterintuitive result here. It may reflect a view that direct payments, however welcome, do not build anything lasting. The provision of AI tools for communities can help find local solutions to pressing challenges. Public services create local infrastructure, local jobs and local expertise. Cash does not. People in these regions may be less interested in receiving AI's dividends than in having AI build and strengthen the systems they depend on.
If AI automates jobs, what do you want the most?
Percentage selecting each option as top preference — by household economic situation
Free public services
Overall
1,036
57%
Western Europe
137
47%
Sub-Saharan Africa
142
59%
South Asia
224
63%
Southeast Asia
82
65%
Middle East & N. Africa
75
61%
Eastern Europe
51
51%
East Asia
100
59%
Latin America
88
55%
North America
124
45%
Oceania
13
62%
0
10
20
30
40
50
60
70
Lower prices on goods
Overall
1,036
23%
Western Europe
137
33%
Sub-Saharan Africa
142
19%
South Asia
224
20%
Southeast Asia
82
24%
Middle East & N. Africa
75
21%
Eastern Europe
51
20%
East Asia
100
19%
Latin America
88
25%
North America
124
24%
Oceania
13
15%
0
10
20
30
40
50
60
70
Cash distributed directly
Overall
1,036
10%
Western Europe
137
13%
Sub-Saharan Africa
142
3%
South Asia
224
4%
Southeast Asia
82
2%
Middle East & N. Africa
75
9%
Eastern Europe
51
22%
East Asia
100
13%
Latin America
88
12%
North America
124
21%
Oceania
13
15%
0
10
20
30
40
50
60
70
AI tools for communities
Overall
1,036
11%
Western Europe
137
7%
Sub-Saharan Africa
142
19%
South Asia
224
13%
Southeast Asia
82
9%
Middle East & N. Africa
75
8%
Eastern Europe
51
8%
East Asia
100
9%
Latin America
88
8%
North America
124
10%
Oceania
13
8%
0
10
20
30
40
50
60
70