A new survey paints a picture of a world increasingly unsure about the current path of artificial intelligence (AI). The Edelman Trust Barometer Flash Poll doesn’t reveal a total backlash against the technology but there is a clear crisis of confidence. According to the report rejection of AI outweighs enthusiasm in several major economies. In the US, 54% of respondents say they reject the growing use of AI, compared with only 24% who embrace it. In Germany, the rejection rate reaches 49%, and in the UK 42%. By contrast, only 10% of respondents in China reject AI, while 46% embrace it; in Brazil, 18% reject and 49% embrace. Global sentiment appears to have split into two decidedly different outlooks.
In the UK, 54% say they reject the growing use of AI. Only 24% embrace it.
The divide becomes even more striking when looked at from a socio-economic perspective. In the UK, 71% of lower-income respondents believe people like them will be left behind by generative AI. That figure is 65% in the US and 59% in Germany. Even among higher earners in the US, nearly half – 47% – fear they personally won’t benefit from AI. Many now worry that AI’s gains will be unevenly distributed and the data suggests they are bracing for disruption rather than opportunity.
The reluctance to use AI appears to be driven by issues around trust and artificial intelligence rather than lack of interest or ability. Among those who use AI less than once a month, 70% of Americans, 57% of Germans and 55% of Britons cite trust issues as the main barrier. In the UK, 33% say they do not feel the need to use AI, with 55% identifying discomfort with data use and protection as their chief concern. Interestingly this concern persists despite the fact that negative personal experiences with AI are rare: just 18% of AI rejecters and 8% of AI embracers say they have had a very bad experience with generative AI. Users aren’t rejecting AI because it has personally harmed them, they’re rejecting it because they don’t trust the intentions, competence or transparency of those deploying it.
This ‘trust gap’ is reinforced by differences across industries. Employees in technology and finance are far more enthusiastic about AI in their work. In the tech sector, 55% embrace its use compared with only 15% who reject it. In food and beverage the picture is reversed with 35% rejecting and only 20% embracing. In transport acceptance dips further with a gap of –15 points between embracers and rejecters. Meanwhile, only 25% of non-managers across markets use AI weekly or more compared with 63% of managers. AI is something that happens to workers rather than with them.
However, the report provides an clue for how trust can be built. Knowledge and trust are the biggest predictors of enthusiasm. People who say AI has improved their ability to understand complex ideas are dramatically more likely to trust it. The report shows that trust in AI among those whose understanding improved was 89% in China, 86% in the US, 73% in Germany and 68% in the UK. Among those who saw no impact on their understanding trust levels fall to 63% in China, 50% in the US, 41% in Germany and 33% in the UK. The gap is significant – 36 to 45 points depending on the country. Direct, positive experience with AI is the single most powerful engine of trust identified in the survey.
For employers the picture is even more concerning. More than half of respondents in Germany (62%), the UK (69%) and the US (70%) believe business leaders are not fully honest about job cuts linked to generative AI. But when employees feel that AI is being used to increase productivity rather than eliminate roles enthusiasm rises. In Brazil, 76% say reassurance around this would increase their enthusiasm; in China, 69%; in the UK and US, 62% and 59% respectively. High-quality training has a similar effect – 73% of Brazilian employees and 66% of Chinese employees say it would boost their enthusiasm, as do 62% of employees in both Germany and the UK, and 57% in the US. The message is that transparency and investment in people are not just good practice they are key for adoption.
The report disrupts any assumption that leaders command authority on AI. When asked whom they trust to tell the truth about generative AI, respondents place “someone like me” at the top in every market: 78% in Brazil, 87% in China, 62% in Germany, 68% in the UK and 68% in the US. CEOs are far less trusted: only 33% in Germany, 32% in the UK and 27% in the US. Government fares even worse with trust levels dropping to 29% in Germany, 28% in the UK and 24% in the US. Peer-to-peer learning and credible expert voices matter far more than top-down messaging.
Despite the polarisation the world has not given up on the potential of artificial intelligence to address global challenges. Majorities in Brazil and China believe generative AI will help solve climate change (76%), lift people out of poverty (57% and 73%) and improve mental health (58% and 74%). In the UK and US optimism is slightly restrained with only 33–42% expecting improvements in those areas. The aspiration for social benefit from AI still exists though – it’s just that it is currently overpowered by concerns that the AI systems being built today won’t be trustworthy enough to deliver those sort of benefits tomorrow.
Those who trust AI are between 4.5 x and 6 x more likely to be comfortable with AI handling major parts of their lives.
The report makes very cleat that trust is not an accessory to AI. Those who trust AI are between 4.5 x and 6 x more likely to be comfortable with AI handling major parts of their lives. At EthicAI, working in AI assurance as we do, the implications for trust and artificial intelligence are clear. Stakeholders globally aren’t resistant to AI for its own sake – they’re resistant to untrustworthy AI. Edelman’s report offers a path to show users how AI helps them – give them control, protect their data, communicate honestly and invite them into the transformation. If those adopting AI commit to that approach the same trust gap that could slow AI’s adoption could become the accelerator to turbo-charge its most valuable future innovations.