Deloitte has doubled down on artificial intelligence by expanding its partnership with AI company Anthropic, rolling out its AI assistant Claude to over 470,000 employees worldwide.
This deployment, Anthropic’s largest enterprise rollout to date, aims to embed AI tools across Deloitte’s various teams, including tailored AI “personas” for disciplines like accounting and software development.
The partnership focuses on regulated industries such as financial services, healthcare, and public sector, combining Claude’s safety-first design with Deloitte’s Trustworthy AI™ framework to promote compliance and transparency.
The announcement comes shortly after Deloitte agreed to refund part of an Australian government contract following errors in a report produced with the help of generative AI. The report contained fabricated citations and references, a result of AI hallucinations—when AI generates false information.
Deloitte has updated the report, disclosed its use of AI tools, and assured that the findings remain valid despite the mistakes.
This situation shows the challenges of adopting generative AI: balancing AI’s productivity gains with the risk of inaccuracies. Despite these issues, Deloitte remains confident in AI’s transformative potential, focusing on scalable, responsible AI deployments that improve productivity, regulatory compliance, and operational efficiency across industries.
Executives from Deloitte and Anthropic emphasise responsible AI as key to the partnership, with plans for specialised certification training for 15,000 Deloitte professionals to support AI integration and client services.
Deloitte aims to lead AI adoption in professional services by fostering in-house expertise and client trust.