The European Union has reached a provisional agreement on the Artificial Intelligence Act, which aims to regulate AI systems being used in the European market. The agreement establishes a risk-based approach, with stricter rules for high-impact AI systems and a ban on certain AI practices deemed unacceptable. The Act also provides for a new governance architecture, including an AI Office within the Commission, and an advisory forum for stakeholders. Penalties for violations of the Act have been set, with caps on administrative fines for SMEs and start-ups (move fast and break things, people!). The Act also includes measures to support innovation and transparency, and it will apply two years after it enters into force. The agreement is considered a historical achievement and a significant step towards setting a global standard for AI regulation.
Is this going to be for the good?
From the article
1. rules on high-impact general-purpose AI models that can cause systemic risk in the future, as well as on high-risk AI systems
2. a revised system of governance with some enforcement powers at EU level
3. extension of the list of prohibitions but with the possibility to use remote biometric identification by law enforcement authorities in public spaces, subject to safeguards
4. better protection of rights through the obligation for deployers of high-risk AI systems to conduct a fundamental rights impact assessment prior to putting an AI system into use.
Looking at it for now there is a lot of feel-good language to parse on what exactly the EU is trying to promote so that's why the exceptions speak volumes about their actual thoughts. The provisional bans for AI so far are cognitive behavioral manipulation, the untargeted scrapping of facial images from the internet or CCTV footage, emotion recognition in the workplace and educational institutions, social scoring, biometric categorisation to infer sensitive data, such as sexual orientation or religious beliefs, and some cases of predictive policing for individuals.
We can safely say that for now, the EU is not interested in creating privatised Big Brother. A win is a win! I can say that this goes in the right direction intuitively when it comes to the prohibitions but now when it comes to positive actions, let's see what is cooking. Small businesses get more benefit of the doubt, so can't say that Big Corp is buddy-buddy with the regulators. Oh wait a minute... the scope exempts military and researchers. What's happening over there? Maybe Big Brother is still being birthed.
The geopolitics of it all is where the spice is. Does Africa follow suit and restrain themselves to the same goals and keep open trade with the EU? Does the USA, UK, and Oceania part of the West follow as well or will they go their separate ways and take advantage of a more unregulated market? They do still have to play within human rights talk which is central to their national images. For non-Western nations where that isn't the case like China, they may go full-on accelerationist in the AI race. I see all of them having military exemptions of course. We little folk better be awake to that fact.