The EU’s voluntary transparency code of practice under the AI Act has more than 200 signatories, so Affiliate Leaders spoke to a few to find out why marketing shops should sign up.
The European Union’s (EU) long anticipated AI Act has come into force earlier this month, which includes the Code of Practice on Transparency of AI-generated Content.
This is a voluntary agreement aimed at helping signatories remain compliant with the AI Act.
Those who sign the code of practice and adhere to it, will satisfy the transparency obligations enforced by the AI Act, and will not need to demonstrate compliance by any other means.
Major players who develop or deploy AI in a range of different industries have signed up to the agreement, including Meta, Microsoft and OpenAI. Meanwhile, Dentsu is the first among the big five advertising holding company to sign the pledge.
Transparency has been a focal point of the AI legislation in the EU, and the code of practice establishes a comprehensive framework for compliance for deployers of AI systems, which generate or manipulate synthetic audio, image, video, or text content.
Layered transparency solutions are encouraged, including watermarking , content provenance mechanisms, metadata, and other features.
Transparency requires collaboration
Many tech platforms have signed up to the agreement, with Section 1 acquiring 81 signatories, and Section 2 gaining 152. These include Google, Amazon, Anthropic and IBM.
Affiliate Leaders approached a number of those who have signed to establish their motivations in joining the code of practice.
Microsoft offered its perspective, in which it outlines a need for greater clarity for audiences viewing AI modified content – and better development of the technologies that make this possible.
The company said: “With today’s text watermarking solutions, effective transparency requires collaboration across the entire AI value chain, including model providers. Advancing solutions that work consistently across the AI stack will depend on continued technical progress, industry collaboration, and real-world implementation experience.”
Effective transparency tools are difficult to build, and information about AI content often gets lost as content is edited, shared, and reused across different platforms, Microsoft points out.
This makes the transparency code of practice an opportunity for “industry partners to advance approaches that are practical, scalable, and effective”, Microsoft concluded.
Other prominent signatories gave very similar reasoning, with Meta pointing to its work through forums like Partnership on AI and the Coalition for Content Provenance and Authenticity (C2PA) as evidence of its ongoing commitment to transparency and developing the technology to correctly identify and label AI content.
OpenAI announced its support for “the code and the ambition behind it”, and explained it helped contribute to the development of the framework. The Sam Altman-company also pointed out it was the first US signatory to the EU’s General-Purpose AI Code of Practice.
On the advertising side, Dentsu is leading the way having pledged to implement the transparency measures when collaborating with clients and partners – particularly in labelling AI generated content when deployed in client work,
Marketing content produced by Dentsu “shapes how people engage with brands and information,” making it all the more important to create a responsible approach to AI generated content.
Content, including operational workflows, technical capabilities, and employee training will be designed with transparency embedded, Dentsu confirmed, adding it will continue to work with regulators.
Other advertising and marketing firms that have signed up to the code include BuzzBrothers, Hobbynote and CW Infinity Marketing; as well as other organisations such as Getty Images, Bulgari, Fastweb, and Lufthansa
