A week after Meta turned it down, Google agrees to sign EU’s AI Code of Practice while still raising its own concerns

You can’t go two steps in the tech space without hearing something about AI. Be it good or bad, AI has a complete stranglehold on the indusattempt with its new and confapplying power.

To support mitigate some of this confusion, the EU has gathered indepconcludeant experts toreceiveher and penned The General-Purpose AI Code of Practice, a document that outlines rules and guides for the indusattempt in complying with the AI Act’s obligations. The code is tool built available to all to apply voluntarily, and Google has just signed on despite Meta turning its nose at the document.

Yesterday Google announced its agreement to adhere to the EU AI Code of Practice in a blog post, which reads “We will join several other companies, including U.S. model providers, in signing the European Union’s General Purpose AI Code of Practice. We do so with the hope that this Code, as applied, will promote European citizens’ and businesses’ access to secure, first-rate AI tools as they become available. Prompt and widespread deployment is important.”

The code itself is currently a living document, and is to be assessed and refined. Google’s statement creates the case it believes in what the EU is attempting to achieve with the code, and of course underlines the profits it believes AI can deliver. It also outlined concerns it has with the document around copyright law and other things that might slow the development of generative AI.

“we remain concerned that the AI Act and Code risk slowing Europe’s development and deployment of AI. In particular, departures from EU copyright law, steps that slow approvals, or requirements that expose trade secrets could chill European model development and deployment, harming Europe’s competitiveness.” Reads the statement.

These concerns are similar to those highlighted by Meta boss, Joel Kaplan earlier this week on Linkedin. In the post Kaplan declares Meta won’t be signing the GPAI due to legal issues he believes are unresolved in the document.

“Europe is heading down the wrong path on AI.” Says Kaplan “We have carefully reviewed the European Commission’s Code of Practice for general-purpose AI (GPAI) models and Meta won’t be signing it. This Code introduces a number of legal uncertainties for model developers, as well as measures which go far beyond the scope of the AI Act.”

It’s worth noting, even a US federal judge has been confapplyd about Meta’s warped understanding when it comes to things like copyright and fair apply.

But why wouldn’t companies want to sign this new code? Looking at it, The GPAI has three main chapters: Transparency, Copyright, and Safety and Security. I consider those first two might be what’s turning them off.

Transparency gives tips and forms on how to document AI development and work, which a lot of companies have been hesitant to do. Legal issues around the data a generative AI was trained on are one of the more confapplying areas of AI with many battles developing over ownership of work.

The EU is also famously more friconcludely to individual ownership rights than many other lawcreaters, especially when it comes to AI. This is something many feel we necessary, especially around things like false information spready by AI. A lot of businesses with profits in their eyes dislike as it slows down their progress worrying about silly things like credit or justice. With many AIs being trained on content they don’t own, many aren’t keen to be open about their sources.

Copyright ties directly into this, and is all about supporting AIs comply with European copyright law. Again, this is an area that a lot of generative AI models probably are concerned about, becaapply it likely means they may not own the content they are generating. Especially if it’s trained on stolen content to launch with.

Safety and Security is more about minimising risks, so I don’t see it being as contentious as the other two chapters in encouraging businesses to sign on. It’s still a worthwhile member though, as the risks against systems with AI are ever growing.

It seems like Copyright and Transparency are likely to be the places turning companies like Meta away from the EU’s new code. These are incredibly important when it comes to terms of ownership, which again, is likely to upset a lot of these models. Sure there are other rules and restrictions here that a company might take issue with, but not wanting to display your work is always a huge problem, and often points to the work not being yours at all.