英国广播公司(BBC)的研究表明，AI聊天机器人在总结新闻报道时，存在事实不准确和扭曲的问题。该研究测试了OpenAI的ChatGPT、Google Gemini、Microsoft Copilot和Perplexity四款AI工具，发现超过一半的AI生成内容存在“某种形式的重大问题”。BBC要求这些AI工具总结100篇BBC新闻文章，记者随后进行审核。结果显示，19%的答案在引用BBC时包含不正确的陈述、数字和日期，13%的BBC引言被“篡改或不存在于原文中”。Gemini的问题最为突出，46%的回答被标记为“准确性存在重大问题”。BBC呼吁科技公司解决AI新闻摘要的准确性问题。

🚨 **事实偏差严重**：研究发现，超过一半的AI生成新闻摘要存在重大问题，包括不准确的陈述、数字和日期错误，以及引言篡改等。

🤖 **具体案例分析**：例如，Gemini错误地声称英国国家医疗服务体系(NHS)“建议人们不要开始吸电子烟”，而NHS实际上建议使用电子烟戒烟。ChatGPT在2024年12月声称Ismail Haniyeh是哈马斯领导人，但他已于2024年7月被暗杀。

⚠️ **Gemini问题突出**：在测试的AI工具中，Gemini的准确性问题最为严重，46%的回答被标记为存在重大问题。

📢 **BBC呼吁合作**：BBC新闻与时事CEO Deborah Turness呼吁科技公司解决AI新闻摘要的准确性问题，并强调AI扭曲的标题可能导致现实世界的重大危害，呼吁新闻行业、科技公司和政府共同努力。

AI chatbots struggle with factual inaccuracies and distortions when summarizing news stories, research from the BBC has found. The study, which examined whether OpenAI’s ChatGPT, Google Gemini, Microsoft Copilot, and Perplexity can accurately summarize news, found more than half of all the AI-generated output had “significant issues of some form.”

As part of the study, the BBC asked ChatGPT, Copilot, Gemini, and Perplexity to provide summaries of 100 BBC news articles, while journalists reviewed their answers. In addition to finding major issues in 51 percent of responses, the BBC found that 19 percent of answers citing the BBC included incorrect statements, numbers, and dates. Meanwhile, 13 percent of quotes from the BBC were “either altered from the original source or not present in the article cited.”

The study highlighted some examples, including Gemini incorrectly stating that the UK’s National Health Service (NHS) “advises people not to start vaping, and recommends that smokers who want to quit should use other methods.” However, the NHS actually recommends vaping to quit smoking. Another example: in December 2024, ChatGPT claimed Ismail Haniyeh was part of Hamas leadership even though he was assassinated in July 2024.

Overall, the study found Gemini’s responses “raised the most concerns,” as 46 percent were “flagged as having significant issues with accuracy.” The Verge reached out to OpenAI, Google, Microsoft, and Perplexity with requests for comment but didn’t immediately hear back.

Last year, the BBC called out Apple’s new AI-powered news summaries for inaccurately rewriting one of its headlines. Apple responded by pausing summaries for news and entertainment apps, as well as making AI notification summaries more distinct from standard notifications.

In a response to the study, Deborah Turness, the CEO of BBC News and Current Affairs, called on tech companies to address issues with inaccuracy. “We live in troubled times, and how long will it be before an AI-distorted headline causes significant real world harm?” Turness wrote. “We’d like other tech companies to hear our concerns, just as Apple did. It’s time for us to work together — the news industry, tech companies — and of course government too has a big role to play here.”