AI ethics research in Denmark
A short guide to places of interest
Why AI and ethics?
At this point in the development and implementation of AI, it is not the same as email. We can not just expect people to learn by themselves, nor have an intuitive understand of the rights and wrongs of using AI. While many features of AI are indeed helpful, they also come with a lot of pitfalls in terms of privacy, trust and errors.
To view AI through an ethical lens, means to take a step back from the engineering perspective of ‘can we?’ to the more thoughtful and socially minded perspective of ‘should we?’.
Ethical AI is all about implications, and so it is a good starting point to look at why we should put a focus on it.
- Risk minimization
An organization with a strong and healthy ethical AI perspective minimizes the risks inherent in adopting such a new and slippery technology. We call it slippery, because the tools are not transparent, and people tend to fall for marketing speak.
- Competitiveness
Regulators will catch up in time, and it is better to be ahead of the pack when the boundaries are drawn, than having to cut back on all initiatives. There is also the perspective user engagement. Actually doing AI in an ethical and sustainable way, will appeal to users who might otherwise be unsure of the benefits of AI.
- Employee satisfaction
To have an organization with a strong ethical core, ensures that good people will want to join. An ethical approach to AI is a core element of an organization’s strategy to get and retain the best people, who want to do good work and do what is right.
- Social responsibility
A commitment to using AI ethically is a commitment to not straining more our already strained societal structures. Principles of ethical AI will help organizations make decisions that provides the growth potential of AI without causing harm in society.
Connection to Denmark
It is a good idea to take a global perspective on how AI strategy is being formulated. This will also provide important insight into how to go about ensuring better ethical AI frameworks as well as better implementation of these through competency building.
Denmark is an advanced digital society, and the digital literacy of the workforce has been fairly high for a long time. The societal structure of Denmark is also very robust, with a focus on the wellbeing for all through the institutions and initiatives of the welfare state.
These two things combined provide a very good foundation to consider how to go about implementing AI in the most ethically sound way.
Places of Interest
Here are some places to look at if you are interested in what the driving forces behind ethical AI in Denmark is.
Center for AI ethics (University of Southern Denmark)
This academic institution use research funds to examine the boundaries of ethics in AI and how to evaluate initiatives. They take the long view, and look at how ethical AI operate on different dimensions such as ethical principles, rights, social-, situated- and cultural practices as well the technological implications embedded in these considerations.
The cross-sectionality of a research institution such as this centre can provide perspectives that purely public or private organizations might be blind to for practical reasons.
Data Ethics Council (Ministry of Internal Affairs and Health )
Denmark was the first country in the world to establish a Data Ethics Council.
They provide research and insights for policy makers around areas such as AI in healthcare and using AI in fighting child abuse. They also provide tools such as conversational games for data ethics in the workplace or toolkit for cross-sharing of data among departments.
National Strategy for Artificial Intelligence (Digital Agency)
The national strategy for AI is meant to act as framework for responsible development and deployment of artificial intelligence.
The first aim of the strategy is to provide a common, ethical foundation for artificial intelligence with humans at the center. The subsequent three aims are to provide a framework for research, industry and public services respectively.
National Strategy for AI (Free PDF)
AI Competency Alliance (Cross-sector Alliance)
The cross-sector initiative The AI Competency Alliance, is an investment and commitment to share best practices of the development and deployment of AI in industry and society.
The initiative has a strong industry bend, and is not writing much about ethical AI. But as a potentially powerful stakeholder, their views on what should be done and their plans for how to do it - and how they navigate ethics, will be an important insight to counter-balance more conservative views.
Competency for ethical AI
When doing research into ethical AI, two points of view emerge, and they are equally important. The first is to develop AI ethically. This entails decisions about what kind of algorithms to use, what input they work with and how output is moderated.1
The second point of view is the competencies of the users (for example public sector employees). What should we expect of the people working with these solutions to provide service to customers and citizens? And just as important, how are we ensuring that they know this?
Here are some of the competencies to be on the look out for.
- Foundational AI knowledge
People might think they understand the implications of AI, just because they know how to use ChatGPT for some tasks. But there is more to it than that. You would benefit enormously from having basic knowledge of best practice cases and edge cases in your industry.
- Critical thinking
People should be able to critically evaluate the output of AI systems. There are many examples of people handing in unedited AI work.
- Moral compass (Ability to reflect)
A key distinction between learning about AI and learning about ethical AI, is to develop an ability to reflect on whether something is good or bad. To reflect is to be able to talk about the considerations, and make careful decisions about what to do with regards to an AI implementation decision.
- Understanding of data
This is related to the first point. The better people understand the kinds of data that is (and can be) used, the better they know what they are dealing with and the pitfalls ahead.
- Understand the users’ perspective
When it comes to ethical AI, we must learn to think about the implications of customers and citizens. The implementation of AI can have very good, and very bad, outcomes for user experience.
For an interesting in-depth perspective on this, I can highly recommend Karen Hao’s Empire of AI↩︎