Today’s digest looks at the ads dominating OpenAI’s trial of ads on ChatGPT, Walmart announcing the phase-out of Vizio accounts on its TVs, and the Trump Administration drawing up new rules following its clash with Anthropic. 

Retail and grocery brands dominate ChatGPT ads trial 

As OpenAI advances with trialling an advertising model on its flagship chatbot, retail and grocery ads are dominating. A month into the trial, app analytics provider Sensor Tower revealed that 44% of ad impressions served on ChatGPT fell under these categories. Retailers involved in the testing include Target, Sephora, and Wayfair, among others. 

This mirrored activity on Google Search during the same time period, which saw the same categories dominate. This makes sense, given that ads are displayed by the chatbot based on the topic of conversation being held and what people are asking about. 

Other leading categories achieved a much lower share of ad impressions served by the chatbot: media and education at 12.6%, travel and entertainment at 10.4%, software and tech at 11.9%, and delivery at 9.1%. 

Walmart to phase out Vizio accounts 

Speaking of retail, Walmart is moving to fully close any remaining gaps between itself and Vizio following its acquisition of the TV maker in 2024. Walmart has announced plans to begin phasing out Vizio accounts: now, when a consumer buys a Vizio TV, they will be asked to sign into or create a Walmart account. Users will be given the option to merge their Vizio account, or delete it.

This change will be limited to newly purchased TVs for the moment, although it will eventually expand to all existing TV owners, according to an email sent by Vizio. 

US draws up new AI guidelines challenging EU regulation 

Elsewhere, following a stand-off between the Pentagon and Anthropic which resulted in both cutting ties and the former striking a new deal with OpenAI to harness its AI technologies, Trump’s administration has drawn up a new set of rules for civilian AI contracts. 

The guidance from the US General Services Administration outlines that AI companies must allow “any lawful” use of their models, effectively granting the US government an irrevocable license to use their technology for any legal purposes. 

As to be expected from the Trump administration, the guidance also mandates that contractors must provide a neutral tool which doesn’t “manipulate responses in favour of ideological dogmas such as diversity, equity, and inclusion”. 

The guidance also mandates that AI companies disclose whether their models have been “modified or configured to comply with any non-US federal government or commercial compliance or regulatory framework”. This clause can be seen to directly challenge compliance with regulatory frameworks within the EU, namely the bloc’s Digital Services Act.