A decisive majority of the British electorate want public policy to regulate AI (Artificial Intelligence). There is consensus on a range of issues relating to AI, including 86% supporting a ban on deepfakes, 60% believing that public policy should prevent superintelligent AI from emerging, and 78% supporting a global watchdog to regulate AI development. This consensus is not polarised along party political lines.
The YouGov poll, commissioned by Control AI and conducted on October 18-19, shows that with regards to both short-term and long-term issues relating to AI, the British electorate reliably and overwhelmingly support regulation.
78% support the establishment of a global watchdog to regulate AI, with only 8% opposing such an organisation.
60% support an international treaty to ban ‘smarter-than-human’ AI
72% believe that public policy should treat AI like a dangerous and powerful technology
79% believe that tech companies should be held liable for harms caused by their technology
86% support a ban on deep-fakes, with only 5% opposing a ban
57% support slowing down the increase in AI capabilities, with only 24% opposing a deceleration
The Need To Regulate
66% of the voters agree that AI is an important issue for public policy, including 63% of Conservative voters and 73% of Labour voters. 74% of voters agree that it should be a goal of public policy to prevent AI from rapidly acquiring superhuman abilities. There is a robust consensus on this across voters of different voting habits, with support being above 70% for all Conservative, Labour, Liberal Democrat, Remain and Leave voters. Remarkably 60% believe that public policy should not just delay but outright prevent AI from acquiring superhuman capabilities.
Immediate Policy Issues
Even with regards to the short-term, non-existential risks that arise from contemporary AI, British voters already show strong support for government regulation. The most overwhelming support for regulations is for a ban on non-consensual deepfakes. 86% of voters support a ban. Additionally, 69% of voters believe that government should be able to regulate contemporary state-of-the-art AI and 70% believe that government should have the power to prohibit AI models from being released.
Unsurprisingly, the British public also has low trust in Big Tech, with only 12% claiming that they have much or little trust in technology CEOs to make decisions in the public interest when it comes to AI. Conspicuously, there is little correlation between voting behaviour and trust in Big Tech - 78% of conservative voters, 81% of Labour voters and 83% of Liberal Democrat voters claim they have little to no trust in technology CEOs. This demonstrates that the antipathy towards Big Tech cannot be explained by political ideology.
69% of polled voters believe that there should be limits on AI models that we don’t currently understand, with only 11% opposed. Considering that a lot of contemporary AI models are characterised as ‘blackbox models’ (meaning that even AI scientists cannot easily understand how they work), these proposed limits would already be applicable to many AI products currently in use.
Specific Regulations
With all possible AI regulations that the poll proposed to voters, they support the regulations. When asked if companies should be required to register large training runs, 74% agreed, 8% disagreed and 18% said they did not know. When asked if AI model creation (as opposed to deployment) should be regulated by policy, 71% agreed, 9% disagreed and 20% said they did not know. In spite of the technical nature of these questions, the vast majority of polled voters chose to affirmatively support these proposed regulations, rather than opting to select the ‘don’t know’ option.
International Regulation & Cooperation
There is overwhelming support for a global watchdog to regulate AI, as well as majorities of all parties’ voters supporting a treaty to ban ‘smarter-than-human’ AI. How voters voted in the 2017 EU referendum does not strongly predict their support for international regulations. 85% of Remain voters and 81% of Leave voters support a global watchdog. When it comes to establishing a treaty to ban superintelligent AI, 60% of Remain voters and 69% of Leave voters support the proposal. Again, this is an example of how broad and non-partisan the public support of AI regulation is. Attitudes towards other international institutions (such as the EU) is not a reliable predictor of attitudes to possible international institutions for regulating AI.
These polling results are a testament to how broad support for AI regulation is among the British public. Public opinion on AI policy is not polarised along party political lines. Neither major political party risks alienating a meaningful portion of its electorate by implementing AI safety policies. There is a broad, demographically and politically diverse mandate among the public in favour of both domestic and international regulations of AI.
Read more about the polling results at i News
Explore the full polling results here
A decisive majority of the British electorate want public policy to regulate AI (Artificial Intelligence). There is consensus on a range of issues relating to AI, including 86% supporting a ban on deepfakes, 60% believing that public policy should prevent superintelligent AI from emerging, and 78% supporting a global watchdog to regulate AI development. This consensus is not polarised along party political lines.
The YouGov poll, commissioned by Control AI and conducted on October 18-19, shows that with regards to both short-term and long-term issues relating to AI, the British electorate reliably and overwhelmingly support regulation.
78% support the establishment of a global watchdog to regulate AI, with only 8% opposing such an organisation.
60% support an international treaty to ban ‘smarter-than-human’ AI
72% believe that public policy should treat AI like a dangerous and powerful technology
79% believe that tech companies should be held liable for harms caused by their technology
86% support a ban on deep-fakes, with only 5% opposing a ban
57% support slowing down the increase in AI capabilities, with only 24% opposing a deceleration
The Need To Regulate
66% of the voters agree that AI is an important issue for public policy, including 63% of Conservative voters and 73% of Labour voters. 74% of voters agree that it should be a goal of public policy to prevent AI from rapidly acquiring superhuman abilities. There is a robust consensus on this across voters of different voting habits, with support being above 70% for all Conservative, Labour, Liberal Democrat, Remain and Leave voters. Remarkably 60% believe that public policy should not just delay but outright prevent AI from acquiring superhuman capabilities.
Immediate Policy Issues
Even with regards to the short-term, non-existential risks that arise from contemporary AI, British voters already show strong support for government regulation. The most overwhelming support for regulations is for a ban on non-consensual deepfakes. 86% of voters support a ban. Additionally, 69% of voters believe that government should be able to regulate contemporary state-of-the-art AI and 70% believe that government should have the power to prohibit AI models from being released.
Unsurprisingly, the British public also has low trust in Big Tech, with only 12% claiming that they have much or little trust in technology CEOs to make decisions in the public interest when it comes to AI. Conspicuously, there is little correlation between voting behaviour and trust in Big Tech - 78% of conservative voters, 81% of Labour voters and 83% of Liberal Democrat voters claim they have little to no trust in technology CEOs. This demonstrates that the antipathy towards Big Tech cannot be explained by political ideology.
69% of polled voters believe that there should be limits on AI models that we don’t currently understand, with only 11% opposed. Considering that a lot of contemporary AI models are characterised as ‘blackbox models’ (meaning that even AI scientists cannot easily understand how they work), these proposed limits would already be applicable to many AI products currently in use.
Specific Regulations
With all possible AI regulations that the poll proposed to voters, they support the regulations. When asked if companies should be required to register large training runs, 74% agreed, 8% disagreed and 18% said they did not know. When asked if AI model creation (as opposed to deployment) should be regulated by policy, 71% agreed, 9% disagreed and 20% said they did not know. In spite of the technical nature of these questions, the vast majority of polled voters chose to affirmatively support these proposed regulations, rather than opting to select the ‘don’t know’ option.
International Regulation & Cooperation
There is overwhelming support for a global watchdog to regulate AI, as well as majorities of all parties’ voters supporting a treaty to ban ‘smarter-than-human’ AI. How voters voted in the 2017 EU referendum does not strongly predict their support for international regulations. 85% of Remain voters and 81% of Leave voters support a global watchdog. When it comes to establishing a treaty to ban superintelligent AI, 60% of Remain voters and 69% of Leave voters support the proposal. Again, this is an example of how broad and non-partisan the public support of AI regulation is. Attitudes towards other international institutions (such as the EU) is not a reliable predictor of attitudes to possible international institutions for regulating AI.
These polling results are a testament to how broad support for AI regulation is among the British public. Public opinion on AI policy is not polarised along party political lines. Neither major political party risks alienating a meaningful portion of its electorate by implementing AI safety policies. There is a broad, demographically and politically diverse mandate among the public in favour of both domestic and international regulations of AI.
Read more about the polling results at i News
Explore the full polling results here
A decisive majority of the British electorate want public policy to regulate AI (Artificial Intelligence). There is consensus on a range of issues relating to AI, including 86% supporting a ban on deepfakes, 60% believing that public policy should prevent superintelligent AI from emerging, and 78% supporting a global watchdog to regulate AI development. This consensus is not polarised along party political lines.
The YouGov poll, commissioned by Control AI and conducted on October 18-19, shows that with regards to both short-term and long-term issues relating to AI, the British electorate reliably and overwhelmingly support regulation.
78% support the establishment of a global watchdog to regulate AI, with only 8% opposing such an organisation.
60% support an international treaty to ban ‘smarter-than-human’ AI
72% believe that public policy should treat AI like a dangerous and powerful technology
79% believe that tech companies should be held liable for harms caused by their technology
86% support a ban on deep-fakes, with only 5% opposing a ban
57% support slowing down the increase in AI capabilities, with only 24% opposing a deceleration
The Need To Regulate
66% of the voters agree that AI is an important issue for public policy, including 63% of Conservative voters and 73% of Labour voters. 74% of voters agree that it should be a goal of public policy to prevent AI from rapidly acquiring superhuman abilities. There is a robust consensus on this across voters of different voting habits, with support being above 70% for all Conservative, Labour, Liberal Democrat, Remain and Leave voters. Remarkably 60% believe that public policy should not just delay but outright prevent AI from acquiring superhuman capabilities.
Immediate Policy Issues
Even with regards to the short-term, non-existential risks that arise from contemporary AI, British voters already show strong support for government regulation. The most overwhelming support for regulations is for a ban on non-consensual deepfakes. 86% of voters support a ban. Additionally, 69% of voters believe that government should be able to regulate contemporary state-of-the-art AI and 70% believe that government should have the power to prohibit AI models from being released.
Unsurprisingly, the British public also has low trust in Big Tech, with only 12% claiming that they have much or little trust in technology CEOs to make decisions in the public interest when it comes to AI. Conspicuously, there is little correlation between voting behaviour and trust in Big Tech - 78% of conservative voters, 81% of Labour voters and 83% of Liberal Democrat voters claim they have little to no trust in technology CEOs. This demonstrates that the antipathy towards Big Tech cannot be explained by political ideology.
69% of polled voters believe that there should be limits on AI models that we don’t currently understand, with only 11% opposed. Considering that a lot of contemporary AI models are characterised as ‘blackbox models’ (meaning that even AI scientists cannot easily understand how they work), these proposed limits would already be applicable to many AI products currently in use.
Specific Regulations
With all possible AI regulations that the poll proposed to voters, they support the regulations. When asked if companies should be required to register large training runs, 74% agreed, 8% disagreed and 18% said they did not know. When asked if AI model creation (as opposed to deployment) should be regulated by policy, 71% agreed, 9% disagreed and 20% said they did not know. In spite of the technical nature of these questions, the vast majority of polled voters chose to affirmatively support these proposed regulations, rather than opting to select the ‘don’t know’ option.
International Regulation & Cooperation
There is overwhelming support for a global watchdog to regulate AI, as well as majorities of all parties’ voters supporting a treaty to ban ‘smarter-than-human’ AI. How voters voted in the 2017 EU referendum does not strongly predict their support for international regulations. 85% of Remain voters and 81% of Leave voters support a global watchdog. When it comes to establishing a treaty to ban superintelligent AI, 60% of Remain voters and 69% of Leave voters support the proposal. Again, this is an example of how broad and non-partisan the public support of AI regulation is. Attitudes towards other international institutions (such as the EU) is not a reliable predictor of attitudes to possible international institutions for regulating AI.
These polling results are a testament to how broad support for AI regulation is among the British public. Public opinion on AI policy is not polarised along party political lines. Neither major political party risks alienating a meaningful portion of its electorate by implementing AI safety policies. There is a broad, demographically and politically diverse mandate among the public in favour of both domestic and international regulations of AI.
Read more about the polling results at i News
Explore the full polling results here
Get Updates
Sign up to our newsletter if you'd like to stay updated on our work,
how you can get involved, and to receive a weekly roundup of the latest AI news.
