New ambitious EU proposal on AI regulation leaves ethical challenges unsolved

The Association of Nordic Engineers (ANE) and the Nordic Privacy Center (NPC) welcome the proposed EU Regulation on a European approach for AI and acknowledge the importance and ambition of this regulatory framework. In the statement we publish today, we draw attention to some of the potential issues of the suggested regulation, which could have…

The Association of Nordic Engineers (ANE) and the Nordic Privacy Center (NPC) welcome the proposed EU Regulation on a European approach for AI and acknowledge the importance and ambition of this regulatory framework. In the statement we publish today, we draw attention to some of the potential issues of the suggested regulation, which could have effects that are contradictory to the ambition and ideals of the legislation.

It will always be difficult to balance any regulation on the application of new technology because of the risk of creating legislative barriers for crucial innovation and business development. On the other hand, a market without any legislation on the use of AI could lead to serious deterioration of personal freedoms and significant loss of workers’ rights, as already seen in some countries around the world.

The statement focuses on the following four key issues:

  • Privacy and data protectionNo “High-risk” AI applications should be deployed without external oversight – both in the private and public sector.
  • Human oversight, responsibilities, and accountabilityThere must be measures included in the proposal outlining how safe spaces for developing the accountability frameworks & the training for human oversight will be created.
  • GovernanceIt is crucial that social partners and third sector organisations are involved in the governance mechanisms. The very people they represent are affected by the deployment of AI systems and therefore, their voices should be heard.
  • Sandboxes and innovationThere should be direct financial support for developers to cover additional costs associated with the regulation. Prioritisation should be given to developing innovative tools to gain transparency into AI systems.  

“Where GPDR has left some holes, including how to limit and protect the use of biometric data, this legislation could potentially move AI forward in a responsible manner. Privacy is not a “means to an end”; it’s a basic human right that we all must cherish. Development of responsible, transparent AI is possible, and hopefully, the EU’s regulations provide a stepping stone forward.”

Cory Robinson, Founder of Nordic Privacy Center

And while it is positive that human oversight and accountability frameworks are mentioned in the proposed legislation – there is a need for clear, stronger measures detailing how the responsibility for this ethical oversight of AI should be distributed between the stakeholders of AI implementation and development.

“We call for more support for human oversight and shared accountability in the development and implementation processes of AI. Without this, individual developers and process managers could find themselves in very difficult ethical dilemmas, where they are left with the responsibility of choosing between ethics and blame or profit and continued work security.”

INESE PODGAISKA, Secretary General for ANE

This was a key theme in ANE’s recent report on ethical dilemmas in AI where engineers from across Europe gave their suggestions and recommendations on what is needed for human oversight to be a realistic process.

ANE and NPC appreciate this very big step forward to regulate the AI ecosystem in Europe and present a way forward for the rest of the world, but as our statement shows there are still many challenges that can only be solved if trade unions and other social partners are invited to be part of the solution.

For further information please contact: Inese Podgaiska (ipo@ida.dk) or Cory Robinson (cory.robinson@liu.se)