Moving to Cambridge to study AI ethics

TL;DR: I'm moving to Cambridge, UK, to study my master's in AI ethics. I'll be back in two years.

If you're reading this, you're likely a work colleague, family member or friend trying to understand why I'm moving to the UK for two years: welcome!

In September 2026, I am taking a two-year sabbatical to study a Master of Studies in AI Ethics and Society at the University of Cambridge. I'm writing this with about three weeks to go, accommodation pending and absolutely no idea how I am going to fit everything I want to take into one suitcase.

How did I get here?

Studying AI ethics was definitely not part of the life plan. I've had the privilege of working at a technologically forward organisation across platform engineering and AI, surrounded by people who have given me opportunities to work on some genuinely interesting problems. I like building things and solving practical problems (which possibly also explains the tiny house), and for a long time I assumed that was more or less the direction I would continue in.

Working more closely with AI complicated that a little. One area I've become particularly interested in is evaluation: essentially, how do you know whether an AI system is actually doing what you think it is doing? This sounds obvious until you try to test one. In conventional software, I can give something an input, define the expected output and check whether the two match. With generative AI, there might be several perfectly good answers, the same input can produce a different response each time, and an answer can be technically correct while still being completely inappropriate. Add agents that can reason through tasks into the mix, and the seemingly simple question of "does this work?" becomes surprisingly difficult.

I had the opportunity through work to explore these problems more deeply while spending time in the US on AI evaluation. I found it fascinating. The technical problems were interesting, but I also noticed that the questions I kept coming back to weren't entirely technical anymore. Who decides what a good outcome looks like? How much confidence do we need before using a system for a particular task, and does that change depending on the industry or consequences? How do you measure safety when the thing you're testing doesn't behave deterministically? At some point, those questions extend beyond another test case.

At around the same time, I was volunteering with the Harding Miller Education Foundation and helping them think through their approach to AI. This offered a very different perspective. Through my work, I've seen the benefit of having engineers, cybersecurity specialists, risk teams, lawyers and governance teams all bring their expertise to decisions about technology. A not-for-profit can face many of the same questions about privacy, bias, reliability and appropriate AI use, but with a much smaller team and budget. I became interested in what "responsible AI" looks like in that context, and how we can make the frameworks being developed around AI genuinely useful to organisations of different sizes.

This eventually became the basis of the research proposal I submitted to Cambridge, looking at AI ethics toolkits and broader policy support for not-for-profits across Australia, the UK and Europe. In hindsight, perhaps this isn't quite the career detour it initially appears to be. Even my undergraduate honours research wandered into data privacy, technology and smart cities. Apparently I've been circling questions about the relationship between technology and society for longer than I realised; AI has simply given me a much larger and considerably messier version of the problem.

The process of applying for a master's

To be frank, applying for master's degrees is a particularly effective way of discovering just how many PDF documents a person can accumulate.

I knew I was interested in questions somewhere between technology, policy and society, but wasn't entirely sure which academic discipline they belonged to. Naturally, this resulted in an unreasonable number of university tabs open at once.

I explored public policy, digital governance, data ethics and AI ethics. I was accepted into the Master of Public Policy at the Hertie School in Berlin, seriously considered Edinburgh's Data and AI Ethics programme, and spent a lot of time looking at Oxford before getting cold feet and not applying at all. Somewhere amongst this, I found Cambridge's MSt in AI Ethics and Society. It was interdisciplinary, part-time and remarkably close to the collection of questions I'd accumulated through work and volunteering. So in February, I gathered yet another collection of PDFs and sent off an application.

Then I was invited to interview.

I expected the Cambridge interview to be terrifying, but to my surprise, it was one of the most enjoyable parts of the process. I genuinely enjoyed talking to the interviewers about the course and their research. Towards the end, they asked whether I had any questions for them, so I asked: "How has teaching a course on AI ethics, an area changing rapidly as we speak, challenged or affirmed your current research hypotheses?"

There was a brief pause where they seemed genuinely surprised by the question (perhaps it was before their morning coffee), and then delighted by it. We ended up having a fascinating conversation about how teaching the course during such a rapidly changing period had influenced their own research and thinking.

I remember leaving the interview less concerned with whether I'd performed well and mostly thinking: I want to have more conversations like that.

Then I got in.

This was obviously the entire point of applying, but there is a slight flaw in applying for ambitious things: occasionally somebody says yes, and then you actually have to rearrange your life around the thing you asked for.

Technically, I could complete most of the degree from Sydney. It would be significantly easier, considerably cheaper and objectively the more practical decision. Unfortunately for me, I got added to the course WhatsApp group.

Candidates started meeting each other, going to conferences, attending talks and debates, and sharing interesting events happening around Cambridge. Someone had produced a film exploring AI and social commentary. A book club was organised for 2.30 am AEST (not my regular reading hour). The course hadn't even properly begun, and there was already this community forming around it. I found myself sitting in Sydney reading the messages and feeling incredibly jealous.

It also made me realise how much I would gain simply by being closer to the broader AI ethics and governance community in the UK and Europe. There is a significant amount of research, policy work and public debate happening there, with universities, think tanks, policymakers and civil society all approaching these questions from different directions. For someone interested in how ideas about responsible AI move from principles into practice, it feels like a particularly interesting place to spend two years.

I realised I didn't just want the degree. I wanted all the marginalia that happens around it: talks unrelated to my dissertation, conferences I wouldn't otherwise attend, conversations after seminars and the chance to learn alongside people approaching AI from philosophy, sociology, politics and disciplines very different from my own.

And, really, I get to study at Cambridge! It still feels slightly ridiculous to say that. If I am going to get my day at Hogwarts, I may as well eke it out to the fullest. I'm under no illusion that every rainy Tuesday in February will feel magical, but this is almost certainly a once-in-a-lifetime opportunity, and I'd like to make the most of it.

A career side quest

I've been trying to find a good way to describe what I'm doing for the next two years, and I've settled on calling it a career side quest.

I'm not stepping away because I've stopped enjoying engineering, nor am I going to Cambridge with a carefully mapped-out plan to emerge in 2028 as an AI ethicist. I quite like being an engineer / product person. If anything, working in technology is what made me curious enough about these questions to want to study them properly.

I'm also incredibly fortunate to work somewhere that has supported me in taking two years away to pursue this. A sabbatical gives me the rare opportunity to follow this particular rabbit hole properly while still having a career I'm genuinely excited to return to.

So, I'll study AI ethics, take long walks, snap an unnecessary number of film photos, meet people I wouldn't have otherwise and eventually write a 15,000-word dissertation on a topical AI ethics problem that will undoubtedly consume my life in 2028. I'm hoping I'll come back with some new ways of thinking about technology and the questions that come with it.

So, why this blog?

A surprising number of people have asked me to share what I'm reading and learning over the next two years, partly out of genuine interest and, in some cases, so they don't have to read all the papers themselves. Fair enough.

So that's what I plan to do here. I'll use this blog to document what I'm learning at Cambridge, unpack some of the research and ideas I come across, and share frameworks, tools and resources I find genuinely useful. My hope is that, over time, it becomes a useful resource for anyone grappling with similar questions about AI's role in society, particularly people trying to turn broad ideas about responsible AI into something practical.

My flight leaves shortly. Before then, there is still work to finish, people to say goodbye to, an alarming amount of life admin and one suitcase that is already looking suspiciously too small.

Stay curious, Elysée