Public Supports AI Safety Summit But Thinks Big Tech Has Too Much Influence
The public are broadly in support of the government's AI Safety Summit (Alamy)
A majority of the public think it is important that the UK hosts the AI Safety Summit, but more than half believe large technology firms hold too much influence over developing AI policy, according to a new poll.
On Wednesday and Thursday, the UK government is hosting the AI Safety Summit which will host delegates from countries from around the world at the historic Bletchley Park site, as well as top tech firms and civil society organisations, with the aim of discussing and understanding the risks posed by frontier AI.
However, figures across the technology and political landscape are still having doubts over whether the landmark AI Safety Summit has the right focus.
A poll carried out by Savanta UK for PoliticsHome surveyed 2,060 people across the UK on their thoughts on the risks and benefits posed by artificial intelligence, and what role they think the landmark summit this week should play in shaping the technology's future.
The findings showed that the public were broadly in support of the summit, with 76 per cent of respondents saying they thought it was important for the UK to host the event, with only 15 per cent saying they did not consider it to be important.
When presented with a variety organisations involved with AI, and asked whether they held too much or too little influence over policy development, 55 per cent said that large technology companies had too much power over the area.
31 per cent said the UK government had too much influence, while only 19 per cent said the same of charities and civil society organisations.
26 per cent thought charities and civil society organisations did not have enough influence, compared to 23 per cent about the UK government and 8 per cent about large tech companies.
On Tuesday, the government published the list of the confirmed attendees to the summit, which will include delegates from 27 countries, and representatives from 40 top tech industry organisations and 46 academic and civil society organisations.
Prime Minister Rishi Sunak will also be hosting an "in conversation" event with controversial X owner Elon Musk following the summit on Thursday, with Musk confirmed to be attending the summit either in person or virtually.
The summit will be largely focused on improving understanding of existential "frontier risks" posed by AI, such as the development of bioweapons, rising unemployment and poverty across the world, and societal unrest.
Asked whether they think the government's focus should be weighted more towards the opportunities or risks presented by AI, 46 per cent of survey respondents said it should be equally focused on both, while 25 per cent thought it should be more focused on opportunities and 18 per cent leaned towards considering the risks.
Overall, 39 per cent of respondents said they thought AI will have a positive impact on UK society and economy, while 24 per cent thought the impact would be negative.
Men were more likely to think it is positive, with 45 per cent considering AI likely to have a positive effect, compared to only 32 per cent of women. Men and women were nearly equally likely to think AI will have a negative effect: 23 per cent of men and 25 per cent of women.
The younger the age group, the more likely they are to think AI will have a positive impact: 50 per cent of 18-34 year olds think AI is positive, compared to 42 per cent of 35-54 year olds and 26 per cent of those aged 55 plus.
52 per cent of those surveyed in London considered AI to bring benefits, compared to only 30 per cent of respondents in Wales. Conservative and Labour 2019 voters were equally likely to view AI as a positive or negative thing.
In the week running up to the summit, the government has released a series of announcements relating to areas where AI can be used, including in the NHS and to aid overseas development.
A number of MPs have told PoliticsHome that they hope that after the summit, there will be a renewed focus on ensuring the UK can harness the potential benefits of the technology.
The survey casts some doubt over whether the public trusts the UK government to ensure the benefits of AI are fully realised and developed, as well as ensuring potential risks are prevented from harming the public.
People were split on whether they trusted the government to ensure the UK benefits from the potential opportunities of AI: 40 per cent said they trusted them, while 48 per cent said they did not.
Meanwhile, 38 per cent said they trusted the government to develop policies to mitigate against the potential risks of the technology, compared to 51 per cent who said they did not trust them to do so.
There is significant concern among the public over the risks that AI could pose for society and the economy.
Potential harms include bias and discrimination, worsening global inequality, negatively impacting democratic elections, damaging employment prospects, and creating new risks around crime and online safety.
Of these risks, the survey found people were most worried about the impact on crime and online safety, with 73 per cent expressing concern; followed closely by employment prospects, which 70 per cent of respondents said concerned them.
Of this set of risks, crime and employment are two of the most immediate risks faced by the public from AI – while issues such as global inequality might be considered a more long term, existential risk such as the issues being discussed at the summit.
The AI white paper, published by the UK government earlier this year, identified six examples of where AI had potential to create harm: on human rights, personal safety, fairness, privacy and agency, societal wellbeing and national security.
Asked which of these examples they felt would be most important for the UK to specifically engage with other countries on at the summit, 31 per cent of the public said national security should be the top issue discussed.
Those who responded to the survey mostly described themselves as understanding what AI is, although only 26 per cent said that they know what it is and would feel confident explaining it to someone else.
65 per cent said they know what it is, but not confidently enough to describe it – suggesting knowledge of AI still remains relatively surface level among the British public.
Much of the controversy surrounding the summit has been around which countries should be invited. Delegates from 27 countries have been listed as attending, with China proving particularly contentious among some hawkish Conservative MPs who have accused the UK government of “failing to think long term” about threats posed by the country.
According to The Times, there is now also tension over whether Israel should have been invited to the summit, with Foreign Office officials suggesting that having Benjamin Netanyahu attend via video link would distract from the summit's aims. Ministers reportedly rejected the suggestion.
The Savanta poll surveyed the public on how important they felt it was for seven of the top economies – the USA, Japan, Germany, France, China, India, and Brazil to attend the summit.
Respondents felt it was most important for the United States to attend, at 78 per cent, while only 63 per cent said it was important that Brazil be involved.
Despite the controversy surrounding China, 73 per cent polled said they considered it to be important that the country attends the summit, with only 14 per cent saying it was not important.
PoliticsHome provides the most comprehensive coverage of UK politics anywhere on the web, offering high quality original reporting and analysis: Subscribe