European Commission presents ethical guidelines for reliable AI
The European Commission has published seven guidelines that should lead to reliable artificial intelligence. For example, AI must respect people’s privacy and diversity of skills. The Commission is now asking for a review of the directives.
Artificial intelligence can contribute a lot to society, according to the European Commission, but also poses challenges in the field of legal and ethical issues. To support AI developers with these issues, the Commission is now publishing the seven guidelines. These would also lead to people gaining more confidence in artificial intelligence, according to the Commission.
According to the Commission, AI systems should promote fair societies by upholding fundamental human rights and by not limiting or misleading human autonomy. Trustworthy AI’s algorithms must be secure and reliable enough so that it can handle errors and inconsistencies. Residents must retain full control over their own data. In addition, data about residents may not be used to harm or discriminate against them.
Furthermore, the AI systems must be traceable according to the guidelines and they must take into account the diversity of human skills and needs. Artificial intelligence should be used to bring about ‘positive, social change’ and should improve sustainability. Finally, according to the Commission, reliable AI should have mechanisms that ensure that AI systems are responsible for their results.
The Commission is now inviting people from industry, researchers and governments to test the guidelines. This is part of a pilot phase, which will start in the summer. The European Commission does not clarify exactly what this pilot will look like. The Commission does say that not only European organizations will participate in the pilot. This is because artificial intelligence knows no national borders. Like-minded countries such as Japan, Canada and Singapore are also invited, the Commission said. An evaluation will take place early next year after the pilot phase.
The guidelines are based on recommendations from the High-Level Expert Group on Artificial Intelligence, a panel of 52 artificial intelligence experts. The publication is part of the European Commission’s AI strategy. With this strategy, public and private investments should grow to at least twenty billion euros annually. This should result in more data becoming available, encouraging European talent and increasing trust in artificial intelligence, according to the Commission.