Next Wednesday will see the last round in a “King Kong meets Godzilla”-style contest between the European Union and the global technology sector over proposed regulations from Brussels to control AI. The opening rounds have been fought by lawyers, lobbyists, and bureaucrats over the monitoring of foundation model AI services such as GPT-4, access to source codes, fines for disobeying the Brussels rulings, and other related topics.
However, EU member states France, Germany, and Italy are known to be opposed to the EU’s proposed rulings and to favor self-legislation by the technology sector, as opposed to being constrained by hard rules dictated by Brussels. French AI company Mistral and Germany’s Aleph Alpha have criticized the EU’s tiered approach to regulating foundation models, defined as those with more than 45 million users.
Given that Europe has, as yet, no significant AI industry of its own, it is hard to see why regulating its nascent AI sector is a matter of such urgency for the EU unless it wishes to extend its authoritarian reach into Silicon Valley and beyond. The EU’s General Data Protection Act (GDPR), which came into force over five years ago, has effectively set the precedent for EU directives from bureaucrats in Brussels impacting companies in regions outside the EU, such as the US. As long as a company offers online services that can be accessed from within the EU, Brussels can enforce its directives with heavy fines.
EU’s AI rulings could be delayed until 2024
A major catalyst behind forcing through the proposed hastily prepared draconian AI rulings would appear to be Western politicians’ quasi-religious belief in the God-like powers with which AI is sometimes credited. US President Joe Biden and UK Prime Minister Rishi Sunak have both subscribed to a widespread notion that AI will soon become superior to humanity and pose a greater potential threat than nuclear Armageddon. A combination of advanced CGI in Sci-fi movies and the Silicon Valley sales hype surrounding services such as ChatGPT appears to have convinced leaders in Washington, London, and Brussels that we are rapidly a kind of ground zero for humanity unless governments act now.
However, there are now growing fears, not just among US-based AI providers but also on the part of the French and German governments, that over-zealous and premature regulation of the nascent AI sector could limit effectiveness. According to sources close to the situation, however, the EU may be prepared to modify its proposals somewhat in line with France and Germany’s wishes.
However, it is thought that opposition from those members of the European Parliament (MEPs) insisting that foundation AI models be regulated could have the effect of temporarily blocking the new rulings. Should this happen, the decision on new AI rulings could be postponed until after the EU elections in June of next year.
