CBA appoints AI researcher Professor Mary-Anne Williams as its Chief AI Scientist.

      The AAAI Fellow based in Sydney has joined the UNSW Business AI Lab to lead a team of Distinguished AI Scientists, marking a new phase in Commonwealth Bank’s development of frontier AI technology. On Monday, the Commonwealth Bank of Australia named Professor Mary-Anne Williams as its inaugural Chief AI Scientist, bringing one of the country’s prominent AI researchers from the University of New South Wales following a thorough global search by the bank.

      Williams will guide CommBank’s group of Distinguished AI Scientists, which is dedicated to machine learning, responsible AI, AI security, and generative AI. She holds the Michael J Crouch Chair for Innovation at UNSW and established the UNSW Business AI Lab while also serving as the deputy director of the UNSW AI Institute. Williams is a Fellow at Stanford, a Fellow of the Association for the Advancement of Artificial Intelligence, features on Robohub’s Top 25 Women in Robotics list, and leads the Sydney-based Social Robotics RoboCup team which secured the world championship in 2019.

      Her research over the past 18 months has focused on how organizations should manage and coordinate fleets of generative AI agents on a large scale, which is the area for which CBA is now hiring her to enhance within the bank. The Chief AI Scientist position is new to CBA, but the broader development surrounding it is not. Ranil Boteju was appointed Chief AI Officer in November 2025 following his role as the data chief at Lloyds Banking Group. Earlier this year, CBA launched a $90 million AI-ready workforce program.

      The bank opened a San Francisco Technology Hub this month to colocate its engineers with AI labs, in addition to its existing Seattle Tech Hub. Furthermore, the company is establishing partnerships directly with Anthropic, AWS, Microsoft, and OpenAI rather than through a systems integrator. The overall impact has been positively recognized externally, with CommBank ranking fourth globally and first in the Asia Pacific on the 2025 Evident AI Index, which assesses bank AI maturity. Boteju stated that Williams brings a unique blend of academic leadership, commercial insight, and profound technical knowledge to accelerate their AI strategy.

      Williams remarked that CBA has ‘developed one of the most sophisticated AI capabilities among banks globally’ and emphasized her aim to focus on ‘enhancing our understanding of the societal impacts of AI and fostering ongoing responsible AI innovation’. The recruitment model mirrors that of major US technology and financial services companies over the past two years, which have been hiring senior AI leaders directly from academic institutions rather than peer banks or research labs. CBA’s announcement underlines this strategy as a deliberate framework.

      The premise is that the segment of the AI technology stack that a large bank cannot delegate is the one needing expertise in frontier research combined with internal organizational authority, and it is easier to recruit individuals with both attributes from a university position than from a bank. The context of an Australian bank making this decision is significant. Australia’s securities regulator, ASIC, now participates in a coordinated effort among financial regulators to oversee frontier AI risks to the banking system, a stance initiated by the Bank of England, the Federal Reserve, and the US Treasury after Anthropic's Mythos cybersecurity model revealed industry-scale zero-day discovery capabilities.

      The upcoming supervisory question pertains to how much understanding of frontier research major Australian banks need to maintain internally to manage AI risks within their own operations. CBA’s conclusion, based on the last nine months, suggests that this understanding exceeds what their foundational-model partners can provide alone.

      The competitive landscape reflects a broader recalibration in AI within the financial services sector. Earlier this month, Anthropic released ten financial services agent templates, incorporated Moody's data into the Claude workspace, and developed distribution via Microsoft 365 and Snowflake. Likewise, SAP, Microsoft, and OpenAI have been promoting various forms of agentic banking tools on a similar schedule. A bank that manages its own team of Distinguished AI Scientists, by its own definition, is preparing for a future where the operationally essential component of its technology stack is the integration layer connecting these external products with the bank’s own data and regulatory framework.

      Williams’s start date and specific reporting line within CommBank have not been disclosed, but her research activities at UNSW will continue in parallel, as is customary for senior academic-to-industry transitions at this level.