Disruptive AI calls for a disruptive legal response

- EN - NL

Generative Artificial Intelligence is advancing at unprecedented speed. Its possibilities seem endless, but so do the dangers. How should we regulate this technology so that the risks are mitigated and the opportunities facilitated? That question is answered by Bart van der Sloot, Tilburg University, in his latest book: Regulating the Synthetic Society.

Experts predict that in five years’ time, more than 90 per cent of all digital content will be wholly or partially AI generated. In such a synthetic society, it may no longer be possible to establish what is real and what is not, Van der Sloot shows. Although they are only in their relative infancy, these technologies can already produce content that is indistinguishable from authentic material. The impact of this new reality on democracy, the judicial system, the functioning of the press, as well as on personal relationships, may well be unprecedented.

In his book Van der Sloot maps the inner workings of ’synthetic technologies’ such as Chat GPT, deepfakes and augmented and virtual reality as well as their uses in the fields of education, health and entertainment. He conceptualises their utilisation for fraud, exploitation and identity-theft. And he explores their deeper effects on the post-truth society, the privatisation of the public sphere, and the loss of individual autonomy and societal trust. 

10 recommendations

In order to ensure that the basic pillars of our society are adequately protected, Van der Sloot concludes that 10 fundamental changes to the current regulatory regime should be considered. 

  1. : There is an urgent need for more information, both with citizens and regulators, about Artificial Intelligence. Market parties should be better informed about how the current regulatory framework applies to them.  
  2. Strategic autonomy : Creating public alternatives to privately owned and critical AI as well as broader export and import restrictions should be considered.
  3. Technological requirements : The need for laying down technical requirements, bans and moratoria on synthetic technologies such as voice cloning should be assessed. 
  4. Precautionary principle : The precautionary principle (if the risks of a technology are unknown, it should not be applied) may need to be transposed to generative AI.
  5. Effectiveness : The effectiveness of an AI application needs to be proven before it can be introduced and applied.
  6. Societal interests : Societal interests and the cumulative effects of synthetic technologies should be accounted for in the regulatory regime. 
  7. Legislative changes : Changes to the legal regime may need to be made, inter alia, to criminalise virtual rape, and address the circulation of deepfake porn.
  8. Supervision : A specialised supervisory authority for AI may need to be set up.  
  9. Cooperation : More cooperation is needed between the regulators and governmental authorities of countries that respect the democratic rule of law.  
  10. Sanctions : The capacity for supervisory authorities to lay down fines and the damages awarded to citizens for harm should be so high that they have a deterrent effect.

version of Regulating the Synthetic Society is now available, and can also be obtained as paperback . It is endorsed by leading experts:

  • ’A captivating read’ Judge Maja Brkan, General Court of the European Union.
  • ’Not to be missed’ Professor Luciano Floridi, Yale University.