2025年生成式AI和新闻报告_62页_5mb

报告摘要

Summary of "Generative AI and News Report 2025: How People Think About AI’s Role in Journalism and Society"

Introduction

This report details the rapid growth in public awareness and use of generative AI, particularly in six countries, alongside public perceptions about its impact on society and news media across Denmark, France, Japan, the UK, Argentina, and the US.

Key Findings

Generative AI Awareness and Use

Awareness and Adoption: By 2025, 90% of the surveyed population had heard of at least one generative AI tool, with ChatGPT being the most recognized. Usage increased dramatically from 40% having used such tools in 2024 to 61% in 2025. Weekly use nearly doubled, from 18% to 34%.
Primary Uses: Information-seeking has surpassed media creation as the main use case. Notable growth in asking questions, researching topics, and using AI for news.
Age Differences: youth are more active users; older age groups use AI less frequently despite broader awareness.
Awareness vs. Trust: High awareness doesn't necessarily translate to trust. Trust is primarily reported for ChatGPT, with lower trust in other tools as few users have heard of them.

Public Views on AI-Generated Search Answers

Integration into Search: AI-generated search answers became common in 2025. Around 54% of respondents across countries reported encountering AI-generated search results, with varying country-specific rates.
Click-Through Behavior: Most users (33%) occasionally click through to sources listed in AI search results, though higher among younger users.
Perceived Trustworthiness: Half of respondents find AI search answers at least somewhat trustworthy. Users trust AI summaries for convenience and quickness but mostly for informational, rather high-stakes topics.

Public Assessment of AI in Different Sectors

Prevalence Perception: The public views AI as already widespread in sectors like search engines, social media (especially Japan and the US), and news media. Fewer see it often used in sciences, healthcare, or education.
Expected Impact: Optimism about AI's benefits varies by sector. Generally, sectors like healthcare and science are seen positively, but news, journalists, and governments are viewed with concern. Trust in AI for personal life is mixed, but societal impact is generally pessimistic.
Variant by Country: Public opinion varies by country, with Japan and Argentina more positive about generative AI than European and US countries.

Public Opinions on AI Use in Journalism and News

Comfort Levels: Citing a 'comfort gap', roughly 12% are comfortable with entirely AI-made news, rising to 21% if human oversight is involved. Use of AI in journalistic tasks (spelling/grammar, translation) correlates with its perceived prevalence.
Creativity and Disclosure: Few regular audience-facing AI features are visible on news websites (e.g., summaries, chatbots). Public trust in AI journalism varies, but those trusting news overall are more open to AI's role.

Conclusion

Generative AI adoption grew significantly, integrated into widely used tools and search platforms. While trust for AI tools is generally low outside ChatGPT, it is moderately trusted for search summary use. In news specifically, optimism around AI's benefits is limited globally, with a distinct preference for human-led journalism. Secure, clear deployment of AI by news organizations, emphasizing human oversight and transparent labeling, is recommended. The data indicates AI's growing societal impact and underscores the importance of balance between efficiency and traditional journalistic integrity.