ARTICLE
27 November 2024

AI Observer | Series 2, Issue 5: The Evolving Role Of AI In Media, Regulation, And Advertising

H
HLK

Contributor

HLK logo
HLK is a global cooperation combining Haseltine Lake Kempner LLP and HL Kempner Partnerschaft mbB and provides a full suite of IP services advising across the entire IPR Lifespan™ in all technical and scientific disciplines. With offices in London, Bristol, Munich, Leeds, Glasgow, and Guangzhou (China), HLK provides IP services across the globe. HLK’s resources and expertise are exclusively dedicated to IP protection: safeguarding the inventions, creative designs, brand identities and other innovations of its clients. HLK advises on the strategy, identification, protection, opposition and appeal, exploitation and enforcement of IP rights, and defends its clients from allegations of infringement by focusing on acquiring competitive advantage for its clients. HLK is privileged to work with some of the most exciting and forward-looking businesses in the world which are at the forefront of innovation and product development in their various spheres.
The latest issue of the AI Observer explores generative AI's impact on podcasting with an AI-generated Sir Michael Parkinson, reviews the upcoming EU AI Act's requirements for general-purpose AI, and highlights Coca-Cola's use of AI for rapid ad production.
United Kingdom Technology

Welcome back to the AI Observer, HLK's newsletter tracking the AI legal landscape.

In the final issue of 2024, we take a deep dive into the controversial world of generative AI and podcasts, focusing on the upcoming podcast Virtually Parkinson, featuring an AI-generated Sir Michael Parkinson. We also delve into the EU's AI Act, which sets out the obligations for providers of these large general-purpose AI models and comes into force in August 2025. Finally, we look at Coca-Cola's ad campaign, where generative AI played a key role in rapidly producing a variety of advertisements.

Did you miss the previous issue of the AI Observer? Check out all our previous issues and other AI updates here.

AI and the future of podcasting: Sir Michael Parkinson reimagined

We take a deep dive into the controversial world of generative AI and podcasts, focusing on the upcoming podcast Virtually Parkinson, featuring an AI-generated Sir Michael Parkinson.

Read the full article here.

Legislation watch

A general-purpose AI system will be able to carry out a wide variety of tasks and integrate with other AI systems. Many people are pretty scared by this, and think it will resemble the out-of-control AI VIKI from I, Robot, but without the attractions of Will Smith to fight it.

The EU has stepped in to allay those fears with the EU AI Act, which sets out the obligations for providers of these large general-purpose AI models and comes into force in August 2025. And the obligations (in Articles 53 to 55 of the EU AI Act) are pretty onerous in terms of openness and documentation. Even more so for general-purpose AI systems over a certain size which are seen as having systemic risk as explained here (we've seen a lot of debate on whether they have chosen the right size, but let's gloss lightly over that).

The EU AI Act then introduces a Code of Practice for general-purpose AI, designed to help the AI giants (such as the likes of Google, OpenAI and Microsoft) get to grips with the EU AI Act requirements. Here is the relevant Art. 56 AI Act on Codes of Practice in a nice searchable version of the Act.

In fact, the AI office of the European Commission, which leads the EU's approach to AI has strategically outsourced this Code of Practice (the "Code") to thirteen unpaid independent and highly prestigious experts. And as explained by two of the experts here, they have set out the following sensible principles in a first draft (which can be downloaded here) shared on 14 November 2024:

  1. Alignment with EU rights and values.
  2. Alignment with the AI Act and international approaches.
  3. Proportionality to risks.
  4. A future-proof approach.
  5. Proportionality to the size of a GPAI model provider.
  6. Support and growth of the AI safety ecosystem.

Whilst not as exciting as Will Smith, these developments may reassure the little man by helping the AI giants comply with their new obligations. The next step is a comprehensive review of the first draft by stakeholders. Watch this space for further updates!

AI application of the month

It's not just national treasures like Sir Michael Parkinson that we've seen AI resurrecting recently. Coca-Cola's AI reboot of the classic 1995 "Holidays are Coming" ad campaign hit the screens last week. Three AI studios (Secret Level, Silverside AI, and Wild Card) worked to create three different ads, using various generative AI models. People love the ads. People hate the ads.

Interestingly, Silverside AI were able to develop and deliver 110 different versions of their ad within just a few days, including 27 skylines customised to the cities in which the ad is to air. Regardless of the creative merits of the ads and of the ethical issues at play, it's impressive to see generative AI successfully used for such high-profile campaigns in such a short amount of time. And humbug to all of you who think November is too early to get started on the festive content!

The content of this article is intended to provide a general guide to the subject matter. Specialist advice should be sought about your specific circumstances.

Mondaq uses cookies on this website. By using our website you agree to our use of cookies as set out in our Privacy Policy.

Learn More