LONDON—British people are more likely than not to think artificial intelligence (AI) will improve their personal lives, careers, and the economy and society as a whole, according to a new nationwide survey conducted by Public First in partnership with the Center for Data Innovation. However, two-thirds of Brits (66 percent) also believe AI will likely increase unemployment.

The nationally representative survey of over 2,000 UK adults tracked shifts in public attitudes toward AI since Public First conducted a similar survey last year. It found that curiosity remains the most common emotion UK residents feel about AI, but feelings of worry, anxiety, and fear have increased.

“Brits are becoming more familiar with AI applications in their everyday lives, but most don’t feel confident that they really understand the technology and there’s little consensus about the longer-term impact it might have on UK society,” said Amy Price, Public First’s Director for Technology Policy. “UK adults are instinctively nervous about a perceived lack of transparency – but quick to support AI adoption when a human is ‘in the loop’ and where the use case is clearly explained. For example, respondents were open to AI helping doctors make diagnoses, provided patients are kept informed and given a choice.”

About one-third of British adults (35 percent) reported using an AI chatbot at work, with 13 percent of full-time workers saying they now do so regularly. Additionally, 73 percent of UK residents say they would support AI-assisted health diagnostics if checked by a human, while more than one-quarter of adults under 35 (28 percent) are interested in having an AI girlfriend or boyfriend.

In terms of AI regulation, majorities of respondents voiced support for various near-term policy actions. Notably, 76 percent favored immediately implementing clear labeling for AI-generated content, and 70 percent favored holding AI companies legally accountable for negative technology outcomes.

Yet a solid majority of respondents (56 percent) believed policymakers should move cautiously before creating new laws and regulations to avoid unintended consequences. The public was divided on whether the approach to AI should prioritize responsible development at the expense of speed (43%) or move quickly to stay at the frontier (39%).

“Government and industry leaders must pay close attention to public sentiment on AI regulation to uphold the UK’s global AI leadership,” said Ayesha Bhatti, a London-based policy analyst with the Center for Data Innovation. “The UK has led by example with initiatives like the AI Safety Summit and establishing the AI Safety Institute, demonstrating a commitment to responsible AI governance and innovation. It’s clear that the public supports responsible AI development yet there is similar support for ensuring that the UK stays at the frontier of AI innovation to remain competitive globally. The decisions the government makes now will define the UK’s trajectory in the AI landscape.”

More key findings:

- 29 percent of British adults expect AI will make things better for themselves personally versus 18 percent who think it will make things worse; 40 percent think it will have no impact.

- A plurality (39%) thinks AI will benefit society as a whole versus 31% who think it will worsen conditions; 13% see no impact.

- Only 31% feel confident explaining how modern AI models work.

- Approximately 13% of full-time workers regularly use LLM-based chatbots; awareness of ChatGPT has notably increased over the past year.

Uses:

- While recognizing its potential to advance science and productivity, only 5% list increasing worker wages among important benefits from AI.

- A majority under age 35 express interest in personal tutors or assistants powered by AI.

- Half learned independently how to use these tools; however, 42% prefer formal training.

- Sixty-five percent support user-choice-based diagnostics but oppose mandatory use.

Concerns:

- Sixty-six percent anticipate job losses due to AI; concern is higher among those with higher education.

- Fears about fake content harming reputations top concerns.

- Over half lack confidence in detecting fake content online; concerns about election manipulation reach 72%.

Future:

- Forty-seven percent believe human-level intelligence could be achieved by the 2030s; views remained stable compared to last year.

- Only twenty-one percent view creating superintelligence positively; forty-percent unaware it's an objective for leading labs.

Policy:

- Majorities endorse policies such as clear labeling of content generated by AI and legal liability for companies.

- Banning new research is opposed more than supported.

- Views on prioritizing technological leadership versus responsible development are divided across demographics.

Read the full report here. (PDF)

Contact: Nicole Hinojosa [email protected]