内容来源：https://techcrunch.com/2025/10/06/deloitte-goes-all-in-on-ai-despite-having-to-issue-a-hefty-refund-for-use-of-ai/

内容总结：

全球知名咨询机构德勤集团近日陷入人工智能应用的"冰火两重天"：在宣布与AI新锐 Anthropic 达成里程碑式企业合作的同时，其因使用AI生成错误内容而向澳大利亚政府退还咨询费的丑闻亦被曝光。

根据《金融时报》报道，德勤于本周一宣布将向全球近50万名员工部署Anthropic开发的Claude人工智能助手，双方还将共同为金融、医疗、公共事业等受监管行业开发合规产品。德勤全球技术生态负责人拉吉特·巴瓦在声明中强调，此次合作将"重塑未来十年的企业运营模式"。

颇具讽刺意味的是，就在同一天，澳大利亚就业与职场关系部披露，德勤因其提交的政府咨询报告中存在AI虚构内容，将退还43.9万澳元合同尾款。该部门去年委托德勤进行的独立评估报告被曝存在多处事实错误，包括引用根本不存在的学术文献。目前修正版报告已重新发布，但事件已引发业界对专业服务机构滥用AI的担忧。

这并非个例。近期芝加哥《太阳时报》因AI虚构书单被迫道歉，亚马逊内部文件显示其AI工具有近三成回答存在事实错误，而Anthropic自家聊天机器人也曾在法律纠纷中生成虚假案例。这些事件共同敲响警钟：在人工智能加速渗透专业领域的当下，如何确保信息准确性已成为全行业面临的重大挑战。

中文翻译：

就在被曝光将因政府委托报告中包含人工智能生成的劣质内容而退还费用的同一天，专业服务与咨询公司德勤宣布与Anthropic达成一项具有里程碑意义的人工智能企业协议。

核心要义在于：德勤与Anthropic的这项合作堪称对其人工智能承诺的全民公投，即便该公司自身也正努力驾驭这项技术。而面临这种挑战的远不止德勤一家。

这项声明的时机颇为微妙——甚至堪称滑稽。据《金融时报》报道，就在德勤宣扬其加强人工智能应用的当天，澳大利亚就业与职场关系部表示，这家咨询公司因其提交的报告中存在人工智能幻觉内容，必须退还相关费用。该部门曾以43.9万澳元委托德勤进行"独立保障审查"，报告于今年早些时候发布。《澳大利亚金融评论》8月披露该审查存在多处错误，包括多次引用不存在的学术报告。修正版报告已于上周上传至政府部门网站。据《金融时报》称，德勤将退还政府合同的尾款。

TechCrunch已联系德勤寻求置评，若获回复将更新报道内容。

德勤周一宣布，计划向其全球近50万名员工全面部署Anthropic的聊天机器人Claude。根据Anthropic官方博客消息，自去年建立合作关系的两家公司计划为金融服务、医疗保健和公共事业等受监管行业开发合规产品与功能。CNBC报道称，德勤还计划创建代表不同部门（包括会计师和软件开发人员）的多种人工智能代理"角色"。

"德勤之所以对Anthropic人工智能平台进行重大投资，是因为我们负责任的人工智能理念高度契合，双方携手将重塑未来十年的企业运营模式。"德勤全球技术与生态系统联盟负责人兰吉特·巴瓦在博客中写道，"Claude持续成为众多客户及我们自身人工智能转型的首选。"

这场企业协议（Anthropic称之为联盟）的财务条款未予披露。

此次合作不仅是Anthropic迄今规模最大的企业部署，更彰显出人工智能如何深度渗透现代生活的各个维度——从职场工具到居家日常咨询。

近月来因使用失实人工智能信息而陷入尴尬境地的，德勤并非孤例。

5月，《芝加哥太阳报》不得不承认其年度暑期书单中混入了人工智能生成的推荐，读者发现部分书名纯属虚构，尽管对应作者真实存在。Business Insider获取的内部文件显示，亚马逊人工智能生产力工具Q Business在上线首年就饱受准确性问题困扰。

Anthropic自身亦曾因使用其聊天机器人Claude产生的幻觉信息而遭受指摘。今年早些时候，这家人工智能研究实验室在与音乐出版商的法律纠纷中使用了AI生成的引用依据，其律师随后公开致歉。

英文来源：

Professional services and consultant firm Deloitte announced a landmark AI enterprise deal with Anthropic the same day it was revealed the company would issue a refund for a government-contracted report that contained inaccurate AI-produced slop.
The upshot: Deloitte’s deal with Anthropic is a referendum on its commitment to AI, even as it grapples with the technology. And Deloitte is not alone in this challenge.
The timing of this announcement is interesting — comical even. On the same day Deloitte touted its increased use of AI, the Australia Department of Employment and Workplace Relations said the consulting company would have to issue a refund for a report it did for the department that included AI hallucinations, the Financial Times reported.
The department had commissioned a A$439,000 “independent assurance review” from Deloitte, which was published earlier this year. The Australian Financial Review reported in August the review had a number of errors, including multiple citations to non-existent academic reports. A corrected version of the review was uploaded to the department’s website last week. Deloitte will repay the final installment of its government contract, the FT reported.
TechCrunch reached out to Deloitte for comment and will update the article if the company responds.
Deloitte announced Monday plans roll out Anthropic’s chatbot Claude to its nearly 500,000 global employees on Monday. Deloitte and Anthropic, which formed a partnership last year, plan to create compliance products and features for regulated industries including financial services, healthcare and public services, according to an Anthropic blog post. Deloitte also plans to create different AI agent “personas” to represent the different departments within the company including accountants and software developers, according to reporting from CNBC.
“Deloitte is making this significant investment in Anthropic’s AI platform because our approach to responsible AI is very aligned, and together we can reshape how enterprises operate over the next decade. Claude continues to be a leading choice for many clients and our own AI transformation,” Ranjit Bawa, global technology and ecosystems and alliances leader, at Deloitte wrote in the blog post.
Join 10k+ tech and VC leaders for growth and connections at Disrupt 2025
Netflix, Box, a16z, ElevenLabs, Wayve, Hugging Face, Elad Gil, Vinod Khosla — just some of the 250+ heavy hitters leading 200+ sessions designed to deliver the insights that fuel startup growth and sharpen your edge. Don’t miss the 20th anniversary of TechCrunch, and a chance to learn from the top voices in tech. Grab your ticket before doors open to save up to $444.
Join 10k+ tech and VC leaders for growth and connections at Disrupt 2025
Netflix, Box, a16z, ElevenLabs, Wayve, Hugging Face, Elad Gil, Vinod Khosla — just some of the 250+ heavy hitters leading 200+ sessions designed to deliver the insights that fuel startup growth and sharpen your edge. Don’t miss a chance to learn from the top voices in tech. Grab your ticket before doors open to save up to $444.
The financial terms of the deal — which Anthropic referred to as an alliance — were not disclosed.
The deal is not only Anthropic’s largest enterprise deployment yet, it also illustrates how AI is embedding itself in every aspect of modern life from tools used at work to casual queries made at home.
Deloitte is not the only company, or individual, getting caught using inaccurate AI-produced information in recent months either.
In May, the Chicago Sun-Times newspaper had to admit that it ran an AI-generated list of books for its annual summer reading list after readers discovered some of the book titles were hallucinated even if the authors were real. An internal document viewed by Business Insider showed Amazon’s AI productivity tool, Q Business, struggled with accuracy in its first year.
Anthropic itself has also been knocked for using AI-hallucinated information from its own chatbot Claude. The AI research lab’s lawyer apologized after the company used an AI-generated citation in a legal dispute with music publishers earlier this year.