In today’s Digest, we look at OpenAI expanding ChatGPT ads with a US-based manager and CPA bidding; Amazon taking aim at supply chain dominance; and publishers suing Meta over AI training copyright claims.

ChatGPT ads manager opens to all US advertisers

OpenAI is accelerating the build-out of its advertising business, rolling out a self-serve ads manager in beta to US advertisers of all sizes. The company has removed its previous USD$50,000 (£39,500) minimum spend requirement to attract smaller brands. The expansion marks a significant shift from its initial closed testing phase, opening ChatGPT advertising to a broader mix of SMBs, startups, global brands, and agency groups.

At the same time, OpenAI is developing more advanced capabilities such as third-party measurement and cost-per-action (CPA) bidding features widely seen as essential for maturing ad platforms, though timelines and partners have yet to be confirmed.

The company is maintaining a cautious rollout, initially limiting ads to lower-risk categories such as consumer goods, travel, and digital services, while signalling plans to expand as its safeguards and compliance systems evolve. 

Amazon targets supply chain

Amazon is expanding its logistics ambitions by opening its distribution, parcel delivery, and fulfilment network to external businesses. The launch of Amazon Supply Chain Services extends capabilities already used by hundreds of thousands of sellers and now targets a broader range of industries, with early adopters including Procter & Gamble, 3M, Lands' End, and American Eagle Outfitters. 

By offering these capabilities as a service mirroring the playbook of Amazon Web Services Amazon is leveraging its scale, data, and delivery speed to capture a larger share of the transport market. Executives say ASCS will offer the same scale, efficiency, and reliability that underpin Amazon’s own operations.

Publishers sue Meta over AI copyright

Meta is facing a new copyright lawsuit in a Manhattan federal court, where major publishers including Elsevier, Cengage, Hachette, Macmillan, and McGraw Hill have accused the tech giant of pirating their books and journal articles to train its Llama AI models. 

The plaintiffs, joined by author Scott Turow, allege that Meta copied millions of copyrighted works without permission, spanning textbooks, scientific papers, and novels, to improve its systems’ ability to respond to user prompts. They are seeking class-action status and unspecified damages, arguing that such large-scale use undermines the rights of creators.

Meta has denied the allegations, maintaining that training AI models on copyrighted material can qualify as fair use.