Ethical Intelligence is an ethics consultancy service that specializes in operationalizing AI ethics across product development and data life cycles.
The interview was conducted in November 2022.
4TU.Ethics Blog: Could you please tell us a little bit about yourself, how you became interested in ethics and technology, and how you ended up as the founder of ethical intelligence?
Olivia Gambelin: I grew up outside of San Francisco, the heart of Silicon Valley. This means I saw a lot of the technology before the rest of the world. We were very much used as guinea pigs in that area, so I was seeing technology at its best and its worst. That’s where my first interest was sparked. When you grow up in that area, it’s also just assumed that you’re going to go into tech. There aren’t really any other options. If you’re in tech, and if you’re good at what you do, being a founder is kind of the mentality.
Later, I studied philosophy in Texas but my fascination for it started in physics. I found theoretical physics to be fascinating and I realized that theoretical physics was basically just philosophy, which I absolutely loved. That’s when I switched to studying philosophy and I got into the world and study of morality and ethics. When I was leaving my undergrad studies, I realized that I love ethics and morality, but I wanted to do something practical.
Shortly after finishing my undergrad, I was exposed to the conversation of Data and AI Ethics. This was around the time that the conversation of ethics was beginning to break out of academia, sneaking its way into the discussion of tech. From then on, I knew that this was the space I wanted to work in. So, I completed a masters in AI Ethics, and set out to work as an AI Ethicist.
I soon discovered that there was a huge gap in the market for practical ethics solutions. Even though there was a lot of talk of ethics, there was no clear direction as to how to practically tackle the bigger challenges. How do we mitigate unethical and harmful consequences of AI? How can we design technology with ethics in mind?
It was this gap which led me to founding Ethical Intelligence. Originally, we were focused on building up research and use cases in how ethics could be practically applied to AI. Since then, Ethical Intelligence has been developing alongside the emerging, and now booming field of AI Ethics.
4TU.Ethics Blog: What exactly is the product of Ethical Intelligence and what is your method for ethics consulting?
Olivia Gambelin: We offer an interim solution that we’ve crafted over time. This year we launched the EI Ethics Board. It essentially supplies companies with experts. We run it as an online platform, we populate the platform with experts within our network, which is a network of over 70 verified professionals in responsible and ethical AI. We train them and their work runs through a central methodology for applying ethics. This provides a lot of structure for something which is usually rather unstructured.
The EI Ethics Board is the first kind of product we’ve really pushed onto the market. And this year we will release some specific blueprints for operationalizing ethical AI, not from a technical standpoint but from an organizational operational standpoint. So, rather than looking at models and datasets we’re looking at the level above – to the people and the management, the governance structure within the organization and how that leads to supporting anything that happens on the technological level. We will focus on a couple of principles such as trust, fairness, and transparency.
4TU.Ethics Blog: Is your methodology based on value sensitive design, or do you combine different approaches?
Olivia Gambelin: Our methodology is based on three basic pillars. The first one is regulation. Of course, simply being compliant doesn’t mean you’re off the hook – but you do need to be compliant as a base level. Thus, the first pillar looks at current regulation and incoming regulations such as the EU AI Act, which is going to make a significant difference.
The second pillar is based upon industry best practices. We’re looking at the operational aspects – what works well when it comes to something like change management or governance of entire teams and culture building within a company.
The third pillar is more concerned with value sensitive design and that’s coming from the literature of principles, best design, and centric design. The combination of those practices brought together are the three pillars that we have based our methodology on.
4TU.Ethics Blog: How do you select the members of the EI Ethics Board?
Olivia Gambelin: There are 79 professionals, mostly academics and independent consultants. These are professionals which are either establishing themselves in a career or are already established. They gravitated towards the expert network because they want to work in responsible and ethical AI. They want to ensure good technological development and have a community around them to be able to have access to be able to work on these kinds of projects.
The expert network has only really worked through word of mouth, and there’s an online application form on our website. If initially successful, we go straight through to interviews. The first round of these is concerned with verifying the skill sets outlined in the application form. Are the skills you’re claiming true? Do you really have this expertise?
The second round of the interview focuses on how good your problem-solving skills are. We have a very strong culture of curiosity and non-judgement within our expert network. We are working with customers that fear what they’re having to talk about. The last thing we want to do is to be judgmental. The approach needs to be one of curiosity – to work on how to solve these issues together. We also look for an aspect of bravery, to be able to face and deal with tough decisions and conversations that will happen throughout the process.
We refer to our expert network when we’re creating a customized ethics board for clients. We look at the company itself, asking first – what is the technology that they are working on? What industry are they operating on? In which country? – and any other specific needs they may have. Once we’ve done this research, we look into the expert network for a combination of someone from a management background, or someone from a hard science technology background. The third person is usually someone aligned with the industry that the company operates in. Sometimes, that last position will change depending on the company, but we try and work on matching preferences of work style and making sure the expertise complement each other.
The reason why we select three different experts is to create that standing relationship with the company, establishing a basis of trust and familiarity necessary for open and honest conversations. But the company also retains access to the entire expert network in case new challenges occur.
Once we’ve set up the experts, the company has access to our platform where they communicate with the board. They manage any open projects or challenges, and this is all kept in one place.
4TU.Ethics Blog: How much impact does an ethics board like this have on a company? Have there been cases of misuse of your ethics board?
Olivia Gambelin: Thankfully we’re at such an early stage in the market, that we’re working with clients in a place where they want ethics. So, this isn’t like a PR stunt for them.
We do recognize that this is a risk though, so we monitor the companies closely, and we require leadership buy-ins, quarterly or monthly, in which the leadership team is required to interact closely with the board.
Part of the board’s focus is to design specific protocols and frameworks which we can check. There’s a couple of checks in place which ensure that there’s no ethics washing.
4TU.Ethics Blog: What could be opportunities or collaboration for our community of researchers, such as PhD students, who work on similar topics of ethical AI?
Olivia Gambelin: The expert network is a way for us to be a very big and very small company all at once. Within the expert network we have different community events that we will organize to connect experts with each other and other academics. Someone might for example be looking for a co-author or use the network to promote some of the research they’re doing. We also try to promote collaboration between experts, so within the network we try and create opportunities for people to connect and feel supported. Besides that, we’re starting research projects ourselves. Instead of only working for companies on specific topics, we’ve been able to identify relevant points for our practice.