Practical tips for companies to take AI ethics into account

  • Posted on: 20 September 2018
  • By: Juho Vaiste

The debate around the themes of AI ethics is heated. We are starting to find the main question marks and ethical concerns, and ideas to solve those. Some of the old debates will continue possibly forever, for example, security vs. privacy, advanced AI and superintelligence, technology philosophical discussion, the correct form of cutting the discrimination out, but that's just the nature of ethics.

There are still good ways to do good in the context of AI ethics and ensuring the responsible approach to the AI development. Some of the solutions are easy; they are like the first steps of AI ethics. They are related to organizational culture, and the core reminder is that "remember ethics".

Various suggestions will follow these first steps. Every organization might find their kind of actions, but aligning the development work, strategic level, and design is my approach. The idea is to take the first steps in each of the fields first, and then deepen the discussion and evaluation by combining perspectives.

 

1. Ensure ethical thinking to be a part of your development work

First step: Accept that your work has moral consequences. It is just that simple. If I understand the developers' mindset at all, the research of AI ethics can be considered 1) nonsense 2) useless 3) distracting. I would suggest another kind of mentality: ethicists are interested in the AI development because its importance is massive. Your work is valued more than ever. 

Further steps: Take the considerations of AI ethics to your developer community. Present the issues from your perspectives and discuss with the other developers. This would be more than valuable for the AI ethics community: we try to understand the technology side as good as possible, but we aren't developers, programmers or data scientists. Spread the word and express your views.

The Star of the Technological Responsibility steps: Build your project on AI ethics. Various tools are coming out to help technology developer to discover and fix the problems of fairness in the algorithmic models. Find your idea, share it, make it open source!

 

2. Plan and formulate strategic guidelines for AI ethics

First step: The first step is to formulate ethical principles or guidelines for the development and implementation of AI technologies. The primary technology companies have done it, and the Finnish Ministry of Economic Affairs and Employment are challenging more companies to join the party.

Further steps: One approach is to empower your developers and designers to think more about AI ethics. The importance of the topic might come from the strategies and guidelines, but the practical changes are planned and executed at the operation level, by the valued designers and developers. Give them time and resources, ask them to focus also on the ethical perspective. 

The Star of the Technological Responsibility steps: The AI technologies are likely to transform our societies. There are worries about social consequences: loneliness, unemployment, losing human touch and meaningfulness from the life. Businesses can have an impact on these concerns: Plan and execute a project to tackle these social changes. Support your local communities. Educate people, for free! For example, a Finnish tech company Futurice is running their social responsibility fund called Chilicorn, what a great way to build your company's impact on the society!
 

3. Learn ethical design

First step: Technology is progressing as a mediatating element: we design and plan it, but it also affects us and change how we see the world. Design has a great ethical responsibility in directing our societies and the world to the right course. Discuss and learn about ethics, talk with the people and the ethicists, work out your value system which introduces the things in our societies that are valuable.

Further steps: There is few design tools and frameworks for the ethical design in technology, but everyone is eager for more and advanced versions of these. Collaborate with the framework designers and teach us others how to use the tools.

The Star of the Technological Responsibility steps: How should we work out the situations when the emerging technologies are run with ethically questionable design? Is the Facebook's addictive notifications and pushing way to force users to accept those an example of ethical technology design? How could we affect this kind of problematics, how can the design community express their opinions and work out the ways of increasing the appreciation of great, ethically aligned design?

 

Let's assure together that we have the control over the technological evolution, progress, and growth. Every level is important, and the multidisciplinary collaboration is essential, also outside the academia.