PhD founders have two superpowers: Lessons from Holistic AI’s Emre Kazim
Emre Kazim is one half of leading AI governance startup Holistic AI. Incubated at Conception X four years ago, the team is based between London and Palo Alto, has recently announced an investment by Mozilla Ventures, and its solutions are now available to Microsoft customers through Azure.
The startup is also a reputable voice in the development of global standards for the use of AI, collaborating with institutions like the Council of Europe and the United Nations to help develop effective approaches towards governing AI systems.
We sat down with Kazim this month to discuss what gave the team an edge, work-life balance as a fast-growing AI startup in Silicon Valley, the superpowers of PhD founders and finding the right investor to share the journey.
The following Q&A has been edited and condensed for clarity.
What’s your grand vision?
This is the question, isn’t it? We’re undergoing an industrial revolution, and the impact on humanity and people is just… Words are superfluous. The core problem we’re solving — of trust in AI systems and ensuring that this kind of huge social, political, cultural and economic change is done in a manner that is responsible — is one of the most important questions for humanity right now.
Tell me more about your technology. Has it changed much since your early days?
We provide risk management for algorithms, so companies adopting and scaling AI systems can be confident that they’ve taken the relevant actions to ensure that those systems behave responsibly. At the core, it’s about creating trust in AI, which isn’t the responsibility of a single entity or person, but depends on ensuring there is clear accountability across the supply chain — from developers to users.
When we first started, we were just trying to solve a very specific problem, looking at how to do a risk assessment of an algorithm for one or two systems. We then started looking at how to do this at scale, so it became about systematic and scalable risk management of algorithms. It was a journey — understanding what the problem we were trying to solve in the real world was and acting accordingly.
How did you become one of the biggest AI success stories in a short amount of time?
If you come from research, there’s a certain level of depth and expertise that is very difficult for people to replicate, and people want to hear what you have to say. Thought leadership is one of the superpowers of anybody coming from our background. Just writing that out and communicating it can have a big impact.
We published papers and blog posts, asked big names to co-author pieces, pitched leading publications and important forums, attended events and met people. We developed a good reputation. And from that, we built out.
You were doing this before the hype. Did that influence you in any way?
We were doing it before it was sexy. As startups and scaleups, we’re in the business of being ahead of everyone else — you’re supposed to see something that people haven’t seen quite yet. It was obvious to us that the use of AI systems in society, business and government would cascade and proliferate. ChatGPT was a huge event, and it accelerated the massive amount of attention, interest and adoption, but it was clear that we were moving in this direction, and it was also very intuitive that you would need governance.
From the start, we were very interested in solving the problem of trust in AI and how to interrogate algorithms — which is an engineering problem — with a view to non-engineers who are using or deploying or being impacted by these systems, including regulators, regular citizens, consumers, C-suite executives.
How many hours do you work per day?
I’ve learnt over time to think about whether what I’m doing is consequential. There are days now when I have a single meeting, but it’s a consequential meeting that may pivot the whole company in one way or another. Am I working when I’m swimming or going for a walk? Probably, yeah, because my mind is still here, but I also don’t want to be part of that thing of saying, ‘Yeah, you’ve got to work 24 hours a day’, and then you’re going to burn out completely. I think it’s important that people achieve some balance and understand that it’s a marathon — that they try to be intelligent in how they work. And it’s different for everyone.
What’s the hardest thing you’ve had to learn quickly while transitioning from PhD student to founder? And how did you learn?
Really, really practical stuff. If you’re doing something no one has ever done before, how do you price it? What value are you adding and how do you translate that into monetary value — without overcharging or undercharging? Also, business operations. You’ve got to manage people, hire, put processes in place. It’s a whole set of novel skills that you’re not trained on as an academic and are going to learn quickly.
Overwhelmingly, what is the superpower of every single person on Conception X or from our background? It’s learning, isn’t it? Learning from others, learning from experience — learning quickly. That’s probably the single most important attribute of anyone in our world.
Tell me about your collaboration with Mozilla Ventures. How did you know they were the right investor for you?
Mozilla was very easy and natural. They were at the forefront of responsible internet and want to be at the forefront of responsible and trustworthy AI — it just made complete sense. If you have the liberty and luxury to choose between different funders, find the best partner.
Why did it make sense to be based in the US?
The US is the right market for everyone — it’s the most vibrant, dynamic, the largest market, and just a great ecosystem to be an entrepreneur. The first time we went to San Francisco was about a year and a half after we started the company — and it was consequential. We walked along the seafront and got to the Golden Gate Bridge. We drove around Cupertino to see Apple, went to Mountain View and saw all the Google buildings. We were tech entrepreneurs and we were in Silicon Valley. That’s something I encourage, getting out there.
How did Conception X help?
It’s the people. You will meet others, network, exchange knowledge and get exposure. You’re going to blink, and in five years’ time, many of those people will be doing phenomenal things.
What’s your advice for PhD founders?
Speak to as many people as you can — don’t over-index on one or two. Speak to as many entrepreneurs as you can because the best advice is stage-appropriate and from people who are in similar situations or just a few steps ahead. That is overwhelmingly the best way to learn.
