May 12 News.KPMG The Global Survey on Trust, Attitudes and Adoption of Artificial Intelligence (2025) was released on May 9th. The study was conducted from November 2024 to January 2025 and covered 48,000 respondents in 47 countries (including China), making it the largest study of its kind.
Core Findings:
- The age of intelligence is in full swing - 66% of respondents use it with high frequency AI83% Recognizing its wide range of benefits
- Trust remains a challenge: only 46% respondents worldwide are willing to trust AI systems
- Call for Regulation Becomes Consensus: 70% Public Calls for National and International AI Governance Framework
- Application Pitfalls Not to be Ignored: 66% User Failed to Verify AI Output Accuracy, 56% Experienced AI-Induced Work Errors
- Chinese Users Remain Cautiously Optimistic: Lead the World in AI Trust, Acceptance, Expectations and Optimism
- ChinacareerAI tool usage in the Middle East is a whopping 931 TP3T, well above the global average (581 TP3T)
1AI has learned from the report that while 66% respondents have formed regular AI usage habits, more than half (58%) still don't consider it trustworthy. AI penetration is significantly higher today than in similar studies conducted in 17 countries prior to the release of the ChatGPT in 2022.Instead, public trust is trending downward and apprehension is on the rise.
More than fifty percent (58%) of employees globally actively use AI tools, with thirty percent (31%) using them weekly, if not daily.China's workplace AI adoption rate is as high as 93%, with half (50%) of the users reaching the level of regular adoption, demonstrating a significant lead.
Eighty percent of global respondents are experiencing the dividends of AI technology: automation of daily tasks, personalized services with precise reach, significant optimization of operating costs, and continued lowering of the technological barrier. However, about 80% of the respondents are concerned about the potential risks of AI.About 40% confessed to having experienced negative impacts -- The problems of alienation of human interaction, frequent cybersecurity breaches, the proliferation of disinformation, inaccurate and biased decision-making outputs, and even the accelerated depreciation of professional skills are particularly prominent.
The report shows that 64% of respondents are concerned that the public nature of elections may be eroded by AI deep-fake content and automated water armies. Chinese respondents were more positive:Up to 91% are optimistic about AI prospectsThe percentage of those who are concerned is only 44%.
Seventy percent of respondents globally call for a stronger AI regulatory governance framework, but only 43% believe that current regulations are sufficiently binding. This perception gap has given rise to strong calls for reform:Global respondents of 87% (86% in China) called for a transnational regulatory systemThe media and social media platforms are urged to establish stricter fact-checking mechanisms to curb AI disinformation. The media and social media platforms are expected to improve their fact-checking mechanisms. The call for collaborative governance between industry and government is also growing.