Thomson Reuters unveils industry Trust in AI Alliance

Thomson Reuters has launched the Trust in AI Alliance, a group that brings together AI researchers and engineers from industry and academia to work on methods for building what it describes as trustworthy agentic AI systems.

Thomson Reuters Labs, the company's innovation research centre, convenes the alliance. Founding participants include senior engineering and product leaders from Anthropic, AWS, Google Cloud and OpenAI, alongside experts from Thomson Reuters.

The group's work centres on AI systems that take more autonomous actions and carry out multi-step tasks. Thomson Reuters said this shift makes questions of safety, accountability and transparency more pressing, particularly in professional domains where decisions can carry legal or regulatory consequences.

Focus areas

Thomson Reuters stated that the alliance will focus on practical collaboration. Participants plan to share insights and identify common challenges across organisations. The group also plans to develop shared approaches for reliable and accountable AI systems.

Thomson Reuters said the alliance will share insights and key themes from each session publicly. It noted that the aim is a broader industry conversation on trustworthy AI.

The alliance's first session will focus on engineering trust into agentic AI systems used in high-stakes professional environments, according to the company.

Thomson Reuters operates across legal, tax and regulatory domains. It said this background provides a practical basis for discussing how AI systems operate in complex professional settings where users may need to explain decisions and defend outcomes.

In a statement on the launch, the company framed the alliance as a response to a change in how organisations think about the risks and assurances associated with advanced AI systems.

"As AI systems become more agentic, building trust in how agents reason, act, and deliver outcomes is essential," said Joel Hron, Chief Technology Officer at Thomson Reuters. "The Trust in AI Alliance brings together the builders at the forefront of this work to align on principles and technical pathways that ensure AI serves people and institutions responsibly, and at pace."

Industry participants

Anthropic said it sees the alliance as focused on operational concerns in settings where AI systems may take more autonomous actions.

"Trust in AI systems is essential as advanced technology takes on more autonomous actions in high-stakes settings and industries," said Scott White, Head of Product, Enterprise, Anthropic. "The Trust in AI Alliance is focused on the practical work of making these systems reliable enough to earn the confidence of the millions of professionals who depend on them."

Google Cloud highlighted the importance of data quality and verification for agentic systems used inside organisations.

"Building trusted agents requires grounding models in 'enterprise truth,' connecting them to the fresh, verifiable data that businesses run on," said Michael Gerstenhaber, Vice President of Product Management for Vertex AI at Google Cloud. "Thomson Reuters efforts to bring the industry together and define shared standards will give organisations the confidence to deploy these intelligent systems in high-stakes environments."

OpenAI said it expects the alliance to address both technical and ethical questions that emerge as AI systems take on broader roles.

"We believe in advancing AI that serves people and organisations responsibly," said Zach Brock, Engineering Lead at OpenAI. "Partnering with Thomson Reuters creates an opportunity to collaborate on the shared technical and ethical questions that will shape AI's long-term role in society."

Reliability and verification

Thomson Reuters said participants will explore approaches to reliability, interpretability and verification. It described these as factors that influence whether advanced AI systems earn and maintain human confidence.

The company positioned Thomson Reuters Labs as the convenor of the initiative. It said the organisation's experience sits at the intersection of technology, human expertise and trust.

Thomson Reuters said it expects the alliance to shape frameworks and standards for the next phase of AI development, as more systems move from producing text and recommendations towards planning actions and executing steps across workflows.

The corporation said the alliance will connect AI researchers, engineers and institutional thought leaders. It said the work will contribute to shared understanding around building confidence in agentic systems.