In today’s Digest, we cover the UK enforcing a ban on junk food TV ads before 9pm, Omnicom Media working with Google on a new insights tool, and OpenAI going all-in on audio with an audio-first device expected to launch this year. 

UK enforces ban on junk food TV ads before 9pm

In an effort to tackle rising rates of childhood obesity, the UK has introduced restrictions on the advertising of junk food. Adverts for products high in fat, salt and sugar are now prohibited from appearing on television before 9pm and are banned outright across online platforms, under measures designed to reduce children’s exposure to unhealthy food marketing.

Although the rules formally begin now, much of the advertising industry has already adjusted, having voluntarily complied since October. There have been restrictions placed on many items often viewed as healthy, such as sandwiches and cereals such as porridge oats.

Omnicom teams up with Google for consumer insights tool 

Omnicom has turned its attention to the rapid transformation of search, as AI reshapes how consumers seek and process information. Unveiled at CES 2026, the company is working with Google on a new initiative intended to give brands a more detailed understanding of search behaviour in an AI-led environment, according to Digiday.

The project, underpinned by research from Joanna O’Connell, chief intelligence officer at Omnicom Media, examines how influence is evolving as search becomes increasingly mediated by generative AI. O’Connell’s work suggests that the traditional information gap between brands and consumers has narrowed significantly, forcing advertisers to rethink how they engage audiences as technology plays a greater role in guiding decisions.

Building on this thinking, Omnicom has developed what it calls the “consumer prompt insights tool”, an agentic product designed to analyse a wide range of search signals and give clients a clearer view of how modern search behaviour reflects consumer needs. The tool is intended to capture the increasingly specific questions people now ask, which reveal far more context and intent than traditional keyword searches.

OpenAI goes all-in on audio

OpenAI is accelerating its push into audio AI as part of a broader strategy that extends well beyond refining how ChatGPT sounds. Reports indicate that the company has consolidated multiple engineering, product and research teams in recent months to rebuild its audio capabilities, laying the groundwork for an audio-first personal device expected to arrive within the next year.

The new audio model is designed to deliver more natural-sounding speech and manage interruptions in a way that more closely mirrors human conversation. The model is also reported to be capable of speaking while a user is talking, a feature that current systems are unable to support.The move mirrors a wider industry shift towards voice-led computing.