AI信任_态度与使用_2025年全球研究（英）_118页_6mb

Trust, Attitudes and Use of Artificial Intelligence: A Global Study 2025

Core Content

This global study, conducted by the University of Melbourne and KPMG International, explores public, employee, and student attitudes toward AI, its use, and the implications for trust, regulation, and governance. It is based on data from over 48,000 respondents across 47 countries, capturing the evolving perceptions of AI since the rise of generative AI tools like ChatGPT in late 2022.

Main Findings

Trust and Acceptance of AI

• Trust remains a challenge: Over half (54%) of people are wary about trusting AI systems.
• Differences by economic group: People in advanced economies are less trusting (39% vs. 57%) and accepting (65% vs. 84%) of AI compared to those in emerging economies.
• Emotional ambivalence: Most people feel both optimistic and worried about AI, with 72% accepting its use.
• Confidence in institutions: Higher confidence is shown in universities, research, and healthcare institutions compared to government and industry.

AI Use and Understanding

• Widespread use: Two in three (66%) intentionally use AI on a regular basis, with three in five reporting they can use AI effectively.
• Knowledge and training gaps: Most (61%) have no AI training, and half report limited knowledge.
• Emerging economies lead: They report higher regular use (80% vs. 58%), training (50% vs. 32%), knowledge (64% vs. 46%), and efficacy (74% vs. 51%) than those in advanced economies.
• Demographic factors: Younger, university-educated, higher-income, and AI-trained individuals report more trust, use, and AI literacy.

AI Benefits and Risks

• Mixed impacts: People report both benefits (efficiency, innovation, decision-making) and risks (cybersecurity, privacy, misinformation, job loss).
• Public ambivalence: 42% believe benefits outweigh risks, 32% believe risks outweigh benefits, and 26% see them as balanced.
• Negative outcomes: 80% are concerned about AI, with 40% having experienced or observed negative impacts.

AI Regulation and Governance

• Strong public mandate for regulation: 70% believe AI regulation is necessary, and 87% want laws to combat AI-generated misinformation.
• Expectations for governance: International laws (76%), national government regulation (69%), and co-regulation with industry (71%) are the most expected forms of AI governance.
• Perceived adequacy of current regulation: Only 43% believe existing laws are adequate, and most people are unaware of AI-related regulations.

AI Adoption in the Workplace

• Employee use: 58% of employees use AI regularly, with a third using it weekly.
• Generative AI dominance: Most employees use free, publicly available tools like ChatGPT.
• Emerging economies lead: 72% of employees in emerging economies use AI regularly, compared to 49% in advanced economies.
• Mixed impacts: Over half report performance benefits, but also note increased workload, reduced human interaction, and compliance risks.

Student Engagement with AI

• High usage: 83% of students regularly use AI for studies, with many using it daily.
• Benefits and risks: Students report benefits like efficiency and personalization, but also concerns about reduced critical thinking and collaboration.
• Institutional support: Only half of students report that their education providers have policies or training for responsible AI use.
• Over-reliance: Many students rely on AI without critically evaluating its output, and some present AI-generated content as their own.

Key Drivers of Trust and Acceptance

• Technical capability: People trust AI’s technical ability more than its societal impact.
• Transparency and accountability: Clear assurance mechanisms (monitoring, human oversight, training) are associated with higher trust.
• Education and training: Higher AI literacy and training are linked to greater trust, use, and critical engagement with AI.

Implications

• Need for responsible AI governance: Organizations must invest in training and policies to ensure ethical and safe AI use.
• Global disparities: Emerging economies show higher trust, adoption, and literacy, suggesting a need for targeted support in advanced economies.
• Balanced regulation: Effective, fit-for-purpose regulation and awareness are essential to ensure AI’s benefits are realized without undermining its potential.
• AI literacy as a strategic capability: Investing in AI literacy is crucial for enabling responsible and effective use of AI across individuals and organizations.

Conclusion

The study highlights the transformative role of AI in society and the workplace, but also the need for greater public understanding, ethical governance, and regulatory alignment. While AI adoption is growing rapidly, trust and responsible use remain critical challenges that require attention from policymakers, industry leaders, and educational institutions.