Public trust in artificial intelligence is facing a significant deficit, hindering its growth and adoption. According to a report by the Tony Blair Institute for Global Change and Ipsos, the lack of trust is the biggest single reason people are shy about using generative AI. This skepticism is widespread, with 56% of non-AI users perceiving AI as a risk to society, compared to 26% of weekly AI users.
The report highlights that familiarity breeds trust, with people who use AI regularly tending to trust it more. Trust increases as individuals experience the benefits and limitations of AI firsthand. However, age and profession play a significant role in shaping attitudes towards AI. Younger individuals are more optimistic about AI, while older generations are more cautious. Professionals in tech are more confident, whereas those in healthcare and education are less so, despite potential job impacts.
To build public trust, governments must demonstrate a commitment to making AI work for everyone, not just a select few. AI implementations in public services should prioritize tangible benefits for real people, not just technical efficiency. Providing users with more control over AI systems, such as data management and model transparency, can also help build trust. Establishing clear governance structures and responsible data management practices is essential for mitigating risks and promoting trust.
The declining public trust in AI has significant implications, including potential geopolitical consequences. Trust in AI varies significantly across regions, with China and developing countries exhibiting higher trust levels compared to Western nations. As AI continues to evolve, addressing the trust deficit will be crucial for its successful adoption and growth.
Ultimately, building public trust in AI requires a multifaceted approach that prioritizes transparency, accountability, and user agency. By working together to address these concerns, we can unlock the full potential of AI and ensure its benefits are shared by all.
