【数据创新中心】公众是如何看待AI的_51页_2mb

Summary of "What Does the Public Think About AI?" (U.S. Poll, 2024)

Introduction and Key Findings

This report analyzes a 2024 survey of U.S. adults to understand public perceptions on AI. About one-third are curious but one-fifth more likely to choose negative emotions like fear. Only 32% feel confident about how modern AI models work, highlighting widespread misunderstanding. Usage is low, with 13% using LLM-based chatbots regularly, though awareness of tools like ChatGPT has nearly doubled due to rapid development. The public is concerned about risks like unemployment (59% think AI will increase unemployment) and misinformation, with 62% worried AI-generated content could manipulate elections. Despite these, many support policies to label AI content and hold companies legally responsible, while opposing new taxes or restrictions.

Opportunities and Benefits

AI is viewed as a major economic driver, potentially boosting productivity and advancing science. However, only 7% believe it will directly increase wages; younger adults show interest in AI personal assistants for education, health, and work, such as tutors and fitness coaches. Benefits like accelerating scientific discoveries are acknowledged, but the public is cautious about job displacement and ethical issues.

Concerns and Risks

The most frequent worry is about information manipulation and deepfakes, with 71% fearful of election manipulation. Other concerns include AI-driven job automation, bias in decision-making, and the potential for profound societal risks like human extinction. Around 46% are not confident detecting fake AI content online, indicating a need for better detection tools and education. Demographic and political divides exist, with Republicans focusing on bias risks and Democrats on personal misuse by criminals or governments.

Future Outlook

Fifty-five percent believe AI could reach human-like intelligence by the end of the 2030s, a slight decrease from last year. Most see AI advancing quickly but disagree on whether the U.S. should prioritize technological leadership or responsible development—43% favor staying at the frontier, while 35% prioritize safety. Concerns about superintelligence are high, with 13% supporting a pause in development due to potential dangers.

Policy Views

A majority supports policies such as requiring AI-generated content to be clearly labeled and making AI companies legally liable, with 66% and 64% endorsing these measures. Notably, policies like imposing new taxes or banning AI research receive negative views, with under 25% supporting increases in funding or restrictions. Cross-demographic support varies, with Republicans emphasizing threats from China, while Democrats worry about ethical and workforce impacts.

National Comparison (U.S. vs. China)

The U.S. public strongly prefers maintaining a technological lead against rivals like China, with just 60% supportive of collaboration. Many fear China's advancements could pose risks to security and competitiveness, though opinions highlight uncertainty and a call for balanced, ethical development to avoid unintended consequences.