Where generative AI formats go next

Generative AI has a huge number of potential applications for the many industries that produce ‘content’ – advertising being one of them – and new formats are in development that hint at the direction of their momentum.

Why it matters

The possibilities of AI in a marketing context are extensive and far from limited to specific formats. These are an increasingly prominent aspect of the possible AI arms race among the major tech groups, and the subject of some of Microsoft Bing’s proposals. Other firms – some with far bigger advertising businesses – are now on manoeuvres.

Google tries to solve its big question

When OpenAI released a public version of its ChatGPT bot late last year, everybody knew that Google had been developing a formidable AI operation, and had been a leader in the research side of the field.

But it had two problems: the potency of the technology on its release to the world, and the implications for its vast search advertising business when the search engine’s output is no longer sponsor-friendly links but prose-based answers.

Now it appears to be integrating the technology more and more into its central search operation, according to reports in TechCrunch, from the company’s Google Marketing Live.

• Context: Adding to last year’s automatically created assets, and drawing from landing pages, the new changes aim to make search ads more contextually relevant to the search query.
• AI for ad creators: In a system likely aimed more toward (the incredibly important) small business segment, Google is also introducing a natural language interface to aid search ad creation.
• Imagery: Photographers beware, as the company’s new Product Studio feature seeks to help sellers create AI-generated product imagery based on existing assets.

Spotify plans voice-based AI ads

Semafor reports that audio giant Spotify is developing generative AI tools that will allow host-read adaptations to ads based on variables like geography. For the moment, this will only take place with the host’s permission and won’t yet be stealing any jobs.

Two questions come to mind: is an AI-adaptable ad that powerful if the data that informs the adaptations might be compromised? The second is bigger: as voices and likenesses filter into generative AI systems – similar to CMOs’ worries about their brands being spoofed – who polices, or limits, their use?

Sourced from TechCrunch, Semafor, WARC

[Image: Dall-e]