Ahead of his keynote presentation opening FStech’s The Future of AI in Financial Services, Doug Robertson, Group Head of Responsible AI Management & Strategy at HSBC speaks exclusively to us about why responsible AI is an accelerator rather than a brake, why accountability can never be handed off to vendors or models, and the foundations firms most often underestimate when trying to scale,

Your keynote argues that responsible AI is an accelerator rather than a brake. For readers who still see governance as a blocker, what convinced you the opposite is true?

If you build a use case first and then ask governance to make it safe, governance will inevitably feel like a brake. But if you build the controls, decision rights, and risk assessment into the way you innovate, they become an enabler of speed.

Think of it as a highway, you can’t have cars travelling 70mph only metres apart without the right surface, lanes, signs, and barriers. It’s possible to go faster because you’ve designed the infrastructure with safety in mind.

The same is true of AI. A well-designed governance framework gives colleagues clarity about what they can do, what they cannot do, what evidence they need to produce and who can make a decision. That removes ambiguity, which is one of the biggest sources of friction in innovation.

So responsible AI is not about saying “no”. It is about creating the confidence to say “yes, and here’s how”. Ultimately that allows an institution to move from hundreds of disconnected experiments to deploying AI at enterprise scale with clear accountability and oversight built in.

You make the point that accountability cannot be delegated to vendors or models and must stay with the institution. In practice, what does that look like inside a bank the size of HSBC?

The starting point is quite simple: using vendor technology does not mean outsourcing accountability for using it responsibly, or the outcomes.

Whether you’re using a vendor or internally developed tools, you need to know what you’re using, why, and verify that effective controls are in place to mitigate risks. In practice, that means having clear ownership of AI throughout the lifecycle.

At HSBC, AI governance is built around this lifecycle approach, including the execution of our Risk Management Framework and defined tollgates overseen by our AI governing bodies.

And there is an important cultural dimension: accountability for good use of AI doesn’t sit with a central team. It is a collective responsibility; the central function’s role is to make that responsibility actionable through clear standards and common capabilities so teams can move faster without reinventing the control wheel.


What are the foundational building blocks – whether governance, infrastructure, skills or decision-making – that firms most often underestimate when trying to scale AI with confidence?

The biggest organisational challenges in scaling AI are often not technical; they’re people and process related; these are easy to underestimate given the interest in the technology itself.

With that said, organisations tend to discover challenges in some of the less glamorous technical foundations: data, identity and access management, legacy infrastructure, monitoring and operational resilience etc. These foundations are increasingly important as organisations scale.

With respect to people, a “build it and they will come” approach doesn’t scale. It’s clear that for the best results you need to invest in relevant training, support employees with adoption, and track impact so AI is embedded into real workflows, not bolted on as a side tool.

Where do you see financial institutions most commonly going wrong in their approach to responsible AI, and what separates the firms getting it right?

There are probably three recurring mistakes.

Firstly, treating responsible AI as principles rather than practice: principles must be translated into engineering, architectural standards and testing, monitoring, and governance requirements.

Second, treating AI solely as a technology risk relegates important questions about business decisions, accountability, and impact and outcomes.

Third, applying a one-size-fits-all governance model: the approach needs to be proportionate to risk, with clear accountability, and a focus on outcomes rather than compliance activity.

Drawing on your career experience, how has the conversation around AI oversight and governance shifted over the past few years, and where do you expect it to go next?

AI oversight, or responsible AI, was often framed around relatively familiar questions of bias, explainability, fairness, etc. Those remain critical, but the conversation has broadened considerably.

As Generative AI matured, new questions were introduced around hallucinations, intellectual property, confidentiality, etc. We improved our answers through operational and technical advances.

With agentic AI, the conversation is shifting from point in time assessments to runtime governance. We’re now discussing how we can trust these systems to operate within defined boundaries with human judgment and clear escalation routes, and focusing more on system behaviour, resilience, and continuous monitoring.

With the EU AI Act and evolving UK guidance reshaping expectations, how should firms balance keeping pace with regulation against the drive to innovate and move quickly?

Establish a global framework and layer jurisdiction-specific requirements on top, rather than creating separate frameworks for each regime.

Regulation is a mandatory baseline but not the definition of responsible AI. The key question is whether you can confidently explain what you’re using, where, why, and how you’re governing and controlling its use effectively.

Standardisation enables speed. A common control framework allows firms to meet evolving regulatory expectations while continuing to innovate at pace.


Doug will deliver the opening keynote presentation at FStech’s The Future of AI in Financial Services, taking place at London Hilton Tower Bridge on October 6. Alongside Doug, the conference will feature an excellent roster of senior speakers from organisations including Lloyds Banking Group, NatWest Group and Standard Chartered. Attendance is free from professionals working within financial institutions. Click here to learn more and to sign up.

---