Clifford Chance

Our report – 'Our Relationship with AI: Friend or Foe – A Global Study' – analyses the responses to a YouGov survey put to 1,000 tech policy experts and influencers across France, Germany, UK, and US, covering a range of technology and AI regulation issues.

Developed in conjunction with Milltown Partners, the report explores overall attitudes to technology and AI, the perceived benefits and risks of AI, as well as respondents' support for different regulatory approaches to AI and their confidence in the likelihood of those approaches succeeding.

Key findings from the report reveal:

Positive attitudes to, and perceived benefits of, AI:

• 66% of respondents perceived AI as likely to have a net positive effect on society and the economy.
• There is strong support for the application of AI in everyday tasks in the private sector and optimism about its ability to transform science (88%), medicine (81%), and elements of business operations for the better (87%).
• Germany was the most optimistic country with 72% agreeing that AI would have a net positive impact, while the US was the least optimistic at 57%.

Concerns about the risks posed by wider adoption of AI:

• There are significant concerns about the ability of companies to handle challenging social and policy questions related to AI, including privacy, algorithmic bias, and civil liberties.
• This reflects a wider pattern of concern about how AI could entrench existing inequalities, benefitting big businesses more than young people or those from minority groups. Just 23% believe AI will likely be positive for underrepresented groups, vs. 36% who believe it will have a negative impact.
• There is particular concern about the use of AI to automate more complex tasks, especially anything that is seen to assess people or prompt action, with some issues polarising the audience. For example, the use of AI for facial recognition by law enforcement divides respondents by geography, with 62% in France expressing trust versus 32% in the US.

Support for different regulatory approaches:

• Industry self-regulation is considered a positive first step by 46% of respondents but is seen as inadequate by itself for the majority. There is strong support for a sector-by-sector regulatory approach (62% support) and an openness to the EU's proposed requirement to register high-risk AI systems with a government or EU-run database (85% support).
• Businesses have failed to convince policy influencers that there is a real trade-off between introducing new introducing new regulation and stifling innovation (44% disagreement), but this is matched by a scepticism about regulators' ability to design and implement effective rules, with only a third confident that they will be able to get it right.