Publication
“Regulating AI is a must”

My reply on SvD Debatt on 9 July 2023 to Mathias Sundin's “Vi behöver inte någon AI-paus” (“We don't need an AI pause”): it is not written in stone that the benefits of AI will outweigh the drawbacks. That is why we need regulation – to set a direction that mitigates harm and creates value for the many, not the few.
Read more
- I agreed with Sundin that AI is not about to wipe us all out – but not that it therefore needs no regulation. The line that bothered me most: “AI har orsakat få verkliga problem” (“AI has caused few real problems”).
- Sundin answered with a final reply on 14 July 2023: “AI-reglering blir broms för viktiga innovationer”.
- I wrote it at the breakfast table on the first day of my summer holiday, while the coffee went cold.
Regulation isn't driven by fear of an AI that wants to wipe us out, but by harm that is already happening: artists whose work is used without consent, innocent people held in custody because of facial recognition, applicants screened out by recruitment tools, patients given the wrong treatment – and the systems' use of energy and water. That is what the EU AI Act addresses. Considerate innovation benefits from clear frameworks.
Links
Skills





















































































































































































































































































































































































