The European Commission announced on Wednesday that it has secured over a hundred signatures from companies that have signed the EU’s Artificial Intelligence (AI) pact.
The AI Act, Endorsed by Hundreds of Firms, Except…
Regrettably, Thierry Breton did not witness the unveiling of his latest project, as the Commissioner resigned due to insurmountable differences with Ursula von der Leyen. The roster of companies participating in the pact includes several big names like Adobe, Google, IBM, HP, Qualcomm, Microsoft, and OpenAI, the creator of ChatGPT. However, two major players are notably absent: Apple and Meta.
While the former has not disclosed the reasons for its non-participation, the latter has been more forthcoming, stating that it hasn’t ruled out joining at a later time. Meta expressed its enthusiasm for the EU’s harmonized rules and is currently focusing on compliance efforts under the AI Act
Meta also added: We should not lose sight of the enormous potential of AI to drive European innovation and enable competition, or else the EU will miss this once-in-a-generation opportunity
.
In its statement, the Commission took the opportunity to highlight the three key elements of this initiative:
Participating companies, under the voluntary commitments of the EU’s AI pact, are encouraged to adopt at least three essential actions:• a governance strategy for AI aimed at fostering the adoption of AI within the organization and working towards future compliance with the AI regulation.
• an inventory of high-risk AI systems to identify those that may be classified as high-risk under the AI regulation
• promoting staff awareness of AI and mastery of AI for ethical and responsible AI development.
What Exactly is the AI Act?
On March 13, the AI Act was signed with overwhelming support from the Parliament (with 523 votes in favor and 46 against). However, reaching this point was not without its challenges, given the numerous debates that have occurred in recent months.
As a reminder, the European Commission initially presented its proposal in April 2021, but the arrival of ChatGPT in November 2022 significantly shook up discussions. OpenAI’s generative AI, capable of writing essays, performing translations in seconds, or creating musical/graphical content just as quickly, added new dimensions to the debates.
The AI Act notably includes a ban—except for security reasons—on real-time remote biometric identification, mass surveillance devices, behavioral manipulation, and social scoring systems (note: citizen rating), similar to the current system in China.
Fierce Debates!
France, which did not hold back during the discussions, eventually agreed but with certain assurances, according to Reuters! Among the conditions set, there was a desire not to hinder the development of this technology, like maintaining a balance between transparency, the preservation of industrial secrets, and administrative burdens (?!…) in the implementation of the obligations imposed on high-risk AI systems
.
Among the points of contention, it is worth noting the threshold definition for the most powerful AI models (ChatGP), which will set particular constraints on some projects. This threshold will also be subject to regular updates in light of technological advances (a practice also adopted by the United States or China).
Several entities, such as SACD and SACEM, have urged the government not to abandon certain principles, such as cultural exceptions and copyright protection. The copyright issues were also among the sensitive topics, with an enhanced obligation for transparency on the data used. However, it remains to be seen how—in practice—this can be controlled and proven!
Hi, my name is Disha and I’m a passionate writer and editor at “Jason Deegan”. With a keen interest in all things tech, I strive to bring you the latest news and updates from the world of high-tech.