If you say that the laws will be written by artificial intelligence, you'd think a dystopia to Black Mirror? Yet, in Dubai, has already started it all. The United Arab Emirates have launched a pioneering initiative to engage the AI in legislative processes, stating that it will be possible to generate drafts regulations in automatic, to assess the impact and simplify the language to make it accessible to all citizens.
An announcement that shakes, it fascinates and makes you think. Truly an algorithm can write the laws? And if so, who's in control?
What is writing in the legislative with the AI and how it works
Write a law is a complex process: it requires legal expertise, the balance between opposing interests, the ability to foresee the consequences. Traditionally, it is an exclusive activity of the human legislator, flanked by lawyers, experts and officials. But the AI is changing the rules of the game.
In Dubai, the idea is to use language models to advanced (similar to ChatGPT) to analyze the existing texts, identify redundancies, to formulate new rules starting from the needs of social, economic or technological. The AI can:
- generate a draft standard is consistent with the existing legal framework;
- suggest changes based on scenarios that are predictive;
- translate the text in simplified language for citizenship;
- simulate the impact of legislation on several sectors and social groups.
The project was announced as part of the strategy, the UAE Coders” and aims to transform Dubai into the first jurisdiction by AI design of the world. Foreign Policy
Artificial intelligence and governance: a pairing risky?
The use of AI in the production of laws opens perspectives extraordinary: more speed, less linguistic ambiguity, control of contradictions regulations. The AI can also help make the law more “neutral” by eliminating certain human prejudices. But is it really so simple?
No. The algorithms are not immune from bias: on the contrary, if trained on legislative texts full of distortions (discrimination, historical, rules, sexist, or exclusionary), will tend to replicate them or amplificarle. And there is more: who controls the model? Who decides on which laws are based. Who decides which parameters to use to say “this is a good idea”?
The danger is that the law to lose its human component, cultural, historical, and become a technical product-calculated, maybe even manipulated by anyone who has access to the algorithm.
Dubai as a global experiment: what changes (and for whom)
The case of Dubai is not an isolated case: in Estonia, Canada, and the United Kingdom are exploring how to integrate the AI into the regulatory process, especially at the local level. But the Uae are the first to make a political project explicit.
For a strong centralized State and pro-technology as Dubai, the AI can be seen as a tool of efficiency. But in contexts of democratic and plural, the use of AI in the writing of laws, raises even more profound: it is compatible with the representation? With the parliamentary debate? With public control?
This initiative can be transformed into a laboratory of the future: a justice of algorithmic automated, but also potentially more accessible. Or it may become a dangerous precedent, if used for the strengthening of authoritarian models, or opaque.
And if in the future the laws were all written by AI?
We are facing a fork in the road: use the AI as a tool of support (that simplifies, analyzes, it helps), or delegate to it the functions of the creative and decision-making fundamentals.
The real challenge will be finding a balance: an AI at the service of human justice, not the other way around. For a law to be legitimate, it should be understood, shared, questionable. And it has to come from a society that is living, not from a statistical forecast.
Conclusion: it serves as an ethics of algorithmic writing
The case of Dubai, we are forced to deal with an awkward question: we want the algorithms to tell us what is right? If, on the one hand, AI can improve the functioning of the institutions, on the other, can become an instrument of power to concentrate, difficult to examine.
To explore these implications, I recommend the article Ethics of Artificial Intelligence: Why it concerns us all, which addresses the theme of the relationship between technology and social justice.
Need a new ethics legislation, taking into account the technology, but don't lose sight of the humanity.