CBA names AI researcher Professor Mary-Anne Williams as Chief AI Scientist

The Sydney-based AAAI Fellow joins the UNSW Business AI Lab to lead a team of Distinguished AI Scientists, in the latest step of Commonwealth Bank’s frontier-AI build-out.

Commonwealth Bank of Australia appointed Professor Mary-Anne Williams as its first Chief AI Scientist on Monday, recruiting one of the country’s best-known AI researchers from the University of New South Wales after what the bank described as an extensive global search.

Williams will lead CommBank’s team of Distinguished AI Scientists, a group focused on machine learning, responsible AI, AI security, and generative AI.

Williams holds the Michael J Crouch Chair for Innovation at UNSW and is the founder of the UNSW Business AI Lab and deputy director of the UNSW AI Institute. She is a Fellow at Stanford and a Fellow of the Association for the Advancement of Artificial Intelligence, sits on Robohub’s Top 25 Women in Robotics list, and is the academic whose Sydney-based Social Robotics RoboCup team won the world championship in 2019.

The 💜 of EU tech

The latest rumblings from the EU tech scene, a story from our wise ol' founder Boris, and some questionable AI art. It's free, every week, in your inbox. Sign up now!

Her published research has, over the past 18 months, focused on how organisations should manage and orchestrate fleets of generative AI agents at scale. That is the subject CBA is now hiring her to extend inside a bank.

The Chief AI Scientist title is new at CBA, but the broader buildout it sits on top of is not. Ranil Boteju joined as Chief AI Officer in November 2025 after a stint as Lloyds Banking Group’s data chief. CBA announced a $90m AI-ready workforce programme in February.

The bank opened a San Francisco Technology Hub earlier this month to put its engineers next to AI labs, alongside an existing Seattle Tech Hub. And the company has been building partnerships with Anthropic, AWS, Microsoft, and OpenAI directly rather than going through a systems integrator.

The cumulative effect has been reflected externally. CommBank ranked fourth globally and first in Asia Pacific on the 2025 Evident AI Index, which scores bank AI maturity. Boteju, in the statement, said Williams brings a rare combination of academic leadership, commercial insight, and deep technical expertise that will help accelerate our AI strategy.

Williams said CBA had ‘built one of the most advanced AI capabilities of any bank in the world’ and that her focus would be on ‘advancing our understanding of the societal implications of AI and supporting continued responsible AI innovation’.

The hiring model is the one that major US technology and financial services firms have been running for two years: recruit senior AI leaders directly from academia rather than from peer banks or foundation labs.

CBA’s announcement notes the approach as a deliberate template import. The bet is that the part of the AI stack a large bank cannot outsource is the part requiring frontier-research literacy paired with internal organisational authority, and that the people with both are easier to recruit from a university chair than from an existing bank seat.

The context inside which an Australian bank is making that bet matters. Australia’s securities regulator, ASIC, is now part of a coordinated group of financial regulators monitoring frontier AI risks to the banking system, a posture that began with the Bank of England, the Federal Reserve, and the US Treasury after Anthropic’s Mythos cybersecurity model demonstrated zero-day discovery capabilities at industry scale.

The supervisory question, framed for the year ahead, is how much frontier research understanding the major Australian banks need to hold internally to manage AI exposure inside their own operations.

CBA’s answer, on the evidence of the past nine months, is that the answer is more than the foundation-model partners alone can supply.

The competitive context is the wider category recalibration in financial services AI. Anthropic shipped ten financial-services agent templates earlier this month, pulled Moody’s data inside the Claude workspace, and built distribution through Microsoft 365 and Snowflake. SAP, Microsoft, and OpenAI have been pushing variants of agentic banking tooling on the same timetable.

A bank that runs its own Distinguished AI Scientists team is, on its own framing, hedging against the day when the operationally critical part of its stack is the integration layer between those external products and the bank’s own data and regulatory perimeter.

Williams’s start date and detailed CommBank reporting line have not been publicly disclosed. Her research base at UNSW will continue in parallel, on the standard pattern for senior academic-to-industry appointments at this level.