Breakthrough Artificial Intelligence calls for Ethics by Design
‘You can't plug in ethics like a USB drive.’
Discussions about the application of Artificial Intelligence (AI) are increasingly about 'Ethics by Design'. What exactly do we understand by that concept, we asked the Dutch specialist in the field, Professor Jeroen van den Hoven of the Technical University Delft. His position: ‘What matters is that you don't just have ethical discussions over a glass of red wine on Friday evening, but that you can also digitally materialize them in the organisation on Monday morning.’
Jeroen van den Hoven is Professor of Ethics and Technology at TU Delft. For more than 30 years, he has been working on ethical and social issues regarding computers and digital technology. He heads the Delft Digital Ethics Centre (www.tudelft.nl/digitalethics ) and is also a permanent member of the advisory board European Group on Ethics (EGE). This body advises the President of the European Commission on ethics in relation to innovation and new technology.
Jeroen, what exactly is Ethics by Design?
‘Ethics by Design is the translation of abstract ethical considerations and moral values into requirements for the design of services, products, artefacts, systems, and architecture. Everyone thinks ethics are very important, but the application in practice is often very unclear. As a result, it can be complicated to transparently account for how your ethical principles have made a difference to a specific service or product. The trick is to translate those abstract values and notions into detailed requirements in a plausible and followable way. That is Ethics by Design: the elaboration of moral values and considerations in terms of requirements. For example, when it comes to the fair nature of an algorithm. Or about mitigating risks around the quality, provenance, and management of data. What matters is that you do not just have ethical discussions over a glass of red wine on Friday evening, but that you can also materialize them in the organisation on Monday morning. From abstract and general to concrete and specific. If you don't think about fundamental values in design early, explicitly, transparently, and systematically, they won't have a place in it either. Or will others take the opportunity to load the technology with their own values and ideas.’
Can you name an example where Ethics by Design is already playing a major role?
‘In my view, the best example comes from the World Health Organisation. Back in 2021, the WHO published on the responsible use of AI in healthcare. This year, it added a report on large language models and generative AI. WHO realises: this is the way forward and Ethics by Design must be part of it. Call it medical ethics 2.0. Developments are moving so fast. Netflix took three-and-a-half years to gain a million users. ChatGPT only needed about five days to achieve that amount of user engagement. If we want to give direction to this absurdly fast-developing technology, we need to get there as early as possible. At the design stage, that is. Society today is largely shaped by digital technology. Crucial sectors and institutions are evolving through services and products of US Big Tech companies, loaded with Silicon Valley values. The European Commission is trying to regain the initiative with far-reaching laws and regulations that put European values first. For example, on competition, digital services, privacy, and AI. We will have to wait and see how successful that turns out.’
What responsibilities does that entail?
‘As ethics become increasingly important, you need to be able to deal with ethical issues productively and constructively as an organisation. Not only in the private, but also in the public sector. Around the discussion about the childcare benefits scandal, for instance, there was a lot of debate in The Hague about what is called the human measure. But what exactly does this human measure mean in practice? What does it mean, for instance, for the data applications of implementing organisations?
Companies and institutions must make this tangible, including testing and validation. Not once every five years, but continuously. In The Hague, many departments and implementing organisations are rightly very interested in this. There is a growing awareness that they need to have the skills to solve ethical questions and problems. Not by reducing them to mere technical, financial, or organisational aspects.’
Do you also see dangers that technological advances can cause?
‘Definitely. For instance, we see people becoming more manipulable, because of deep fakes, among other things, which create polarisation in society. Especially at election time, but also outside it, there are attempts to undermine democracy everywhere. These are obviously major threats to everything we hold dear, even though we have enshrined it in the constitution. Companies are asked to be on the right side and understand what that means for their business. For instance, in their dealings with price targeting, price gauging, and flexible pricing. And other AI tools that allow you to squeeze the maximum amount of money out of a consumer. That goes a lot further than a simple seduction technique like candy at the checkout.’
How will Ethics by Design evolve?
‘It will evolve into an important tool in the toolbox of digital technology governance. Of course, we have the existing laws and regulations and a whole arsenal of legal tools. Ethics by Design adds to that. It has yet to find its place, also as part of corporate governance in organisations. That might be a struggle, because of course you cannot plug in ethics like a USB drive. We will have to specify what place ethics has in the organisational chart and how it relates to existing functions. Think of officers dealing with cybersecurity, data protection, and public affairs. That discussion will eventually culminate in the development of standards, norms, and best practices.’
What type of people need to get involved if there is to be a link to business?
‘It is a very multidisciplinary area of interest. Future courses in business administration or public administration, with an emphasis on digital and ethics, will start providing those competences. Ethics by Design will become a natural part of such courses. You need to understand technology, but also philosophy and relevant law. And you need to be conceptually flexible combined with a philosophical mindset.’
‘Digital climate change is imminent if we are not careful.’
Do you have any concluding thoughts?
‘Digital climate change is imminent if we are not careful. Our social, cultural, and personal experience will change dramatically as a result of what Big Tech is pouring out on us in terms of AI. This is going to bring about fundamental changes, only a fraction of which we currently understand. We see only the tip of a very rapidly melting iceberg. That, by the way, is no reason to stand idly by. The lessons of climate change also apply here. As a company, have you simply continued pumping oil for 40 years, or have you asked yourself: what can I do differently? So, the question “Which side are you on?” does become a very relevant one in the coming years.’