• by Laurie Sullivan , Staff Writer @lauriesullivan, May 29, 2024

There is no shortage of data from publishing companies to train large language models (LLMs) for OpenAI. 

Years ago, advertising insiders called data the internet's gold, but the holy grail at the time was to find a method to electronically facilitate the acquisition and the sale of third-party data similar to a stock-trading desk at NASDAQ. 

Now, for companies like Google, Microsoft, and OpenAI, the holy grail has become licensing all types of data -- from readership and website user data from publishers, to social media and search -- so they can custom train their artificial intelligence (AI) models.

The Atlantic, one of the oldest magazines in the United States, has been publishing content since 1857. Vox Media, one of the nation's largest digital media holding companies, is parent to the New York Magazine brand, which has been publishing since 1968.

On Wednesday the two publishers announced separate licensing and product deals with ChatGPT parent OpenAI. These multi-year agreements allow OpenAI to license the publishers' archived content to train its AI models.

One of the most interesting pieces of news from The Atlantic revolves around an experimental microsite called Atlantic Labs that the publisher will use to determine how AI can aid in the development of new products and features to better serve journalism and readers.

The Atlantic also described in the blog post how its articles will become searchable and discoverable in OpenAI’s products, including ChatGPT.

As a partner, The Atlantic will help OpenAI “shape how news is surfaced and presented in future real-time discovery products.” Queries that serve The Atlantic will include attribution and a link to read the full article the publisher’s website.

A spokesperson for The Atlantic said the publisher has a number of AI experts, members of our executive team led by the CEO.

Vox Media will also use OpenAI technology for Forte, its first-party data platform that launched in 2019. The company built the platform to drive outcomes for advertisers, with its targeted impressions earning 2x the performance compared to alternative data sources, it explained in a blog post on Wednesday.

OpenAI said it will include citations when referencing response to a user query and will link out to the relevant article. Sort of like a search engine. (Hum.)

Vox Media believes this partnership will boost the company’s ability to innovate with generative artificial intelligence (GAI), and as part of the agreement, it will work with OpenAI to build audience-facing and internal applications and capabilities.

Discovery and recommendations will become a focus for Vox Media, as well as improving its affiliate commerce product, The Strategist Gift Scout. The produce acts as a search tool, matching shoppers with endorsed gift using OpenAI tools.