UBS has made a strong AI governance move
UBS has made a strong AI governance move – practically, financially and culturally. It’s not the first company to take this kind of proactive stance, but it’s certainly a blueprint for other companies trying to define their policies around artificial intelligence.
So far this decade, business journalists and commentators have been noting different companies’ reactions to AI. Some firms – many in the legal and PR sectors – were quick to ban its use by employees. Swiss multinational investment bank UBS has gone down the opposite road, recently announcing that all new graduates and interns in its global banking and markets divisions will, from 2027, need to demonstrate how they can use AI to “improve outcomes and efficiency”.
The Financial Times has labelled the bank as “one of the first major financial institutions” to make such a declaration, which essentially gives AI proficiency equal importance to other more traditional requirements, like a university degree with high marks.
What has UBS said?
“We continuously review our recruitment criteria to ensure they reflect the skills we need,” the organisation told FT in September 2026.
“AI literacy complements, rather than replaces, the academic, analytical and interpersonal skills that remain core to our approach to hiring talent.”
Why is this a good AI governance move?
The recognition that AI is a complementary tool rather than a replacement for human workers is a common, yet critical first point in why UBS is making a good AI governance call. Despite the noise and endless commentary, the position recognises AI’s practical place in business going forward – one that balances innovation, productivity and risk management.
The kicker, though, is in the drive to “improve outcomes and efficiency”. Newcomers aren’t asked to simply show that they can use AI; they’re asked to show how they can use AI for the benefit of the business. That’s a very powerful distinction.
On a practical level, it speaks for itself. The business benefits as AI takes hold in the workplace. On a cultural level, it’s incredibly important: the policy obviously comes from the board and management level, but it’s being implemented among employees who are just starting, which is a solid method of ensuring that the right attitude to AI persists for the entirety of a worker’s tenure with the business.
Ultimately, any work with AI must be justified. It has to lead to positive results, not just tick-box exercises or work done for appearances’ sake.
It’s said in some sporting circles that practice doesn’t automatically make perfect; it only makes “permanent”. If you do the wrong thing in pursuit of the right goals, you may actually be pursuing wrong goals. There are echoes of that in this debate around AI. Simply requesting that workers can use it – without further clarification or metrics – might look like a nice idea on the surface, but it has no substance. It will likely deliver less in the long term, which, in itself, is a strategic risk.
Embracing AI means thoroughly discussing what success looks like and ensuring those benchmarks are achieved in reality. If not, why not, and what can you learn from it?
What does that look like in practice?
Anyone claiming to be proficient enough in AI for positive business outcomes must be able to do the following:
- Translate corporate challenges into coherent AI-based research, and translate AI results into coherent business action points.
- Process data correctly. This goes for both sides of the AI journey: competent individuals can structure data before AI sees it, so that it will be read properly, and check/understand subsequent data that AI has produced.
- Prompt far beyond the “beginniner” level. They can write precise, relevant instructions that can be built into workflows if and when the company needs it.
- Understand the internal and external regulations around AI use and ensure their own usage never breaks the rules, especially where private data is concerned.
The pivotal lesson for boards
The crucial point is that all of the above practical elements matter just as much to executives and the board as they do for incoming graduates and interns. Not everyone has to understand the unique tech specs of AI, but the level of fluency above is quite possible to achieve with just a little training, and realistically, it’s essential at all levels of the business.
Remember, you’re not just using AI because stakeholders want you to, because everyone else is, or because you think your company will miss out without it; all of those reasons carry no quantified success metrics, while raising significant risk of overinvestment and underperformance.
You’re using AI to deliver practical benefits. You can’t do that without trained personel at all levels of your business. If your board doesn’t have anyone with this kind of AI-literacy, it might be time to think about acquiring it.