The Credit They Deserve: As AI Expands Into Newsrooms, Are Publishers Seeing Better Attributions?

Recently, Axios announced that it would be expanding its local newsrooms to 34 in total. 

And while local-news expansion is always welcome in the publishing world, it’s the publisher’s partner that makes this expansion particularly noteworthy: OpenAI.

Yes, the AI company is funding the local-news expansion to four cities, “as part of a broader content-sharing and technology deal,” Axios says. And while Axios CEO Jim VandeHei says OpenAI will not be used for reporting in Pittsburgh, Kansas City, Boulder, or Huntsville, it will “help build a system for creation, distribution, and monetization of our journalism.”

“ChatGPT will use Axios journalism to answer user queries with attributed summaries, quotes and links to Axios stories,” writes Sara Fischer of Axios, which reaches more than 2 million subscribers with its newsletters. “Axios can access OpenAI’s technology to build its own AI products, processes, and systems.”

OpenAI’s newsroom funding is just the latest evolution in a number of wide-ranging agreements with publishers and media organizations that allow them to leverage AI technology for their own use. In return, as OpenAI highlights, the “partnerships with publishers like News Corp, The Atlantic, Vox Media, Prisa Media, Condé Nast, Hearst, and Dotdash Meredith enable ChatGPT search⁠ to now feature select summaries and excerpts from trusted media outlets with clear citations and direct links to original sources.”

That search value, however, comes into question when considering a study from The Tow Center, published late last year in the Columbia Journalism Review, that says “publishers face the risk of their content being misattributed or misrepresented regardless of whether they allow OpenAI’s crawlers.”

In asking ChatGPT to identify the sources of 200 quotes from a group of 20 publications — a group “representing a mix of those who have deals with OpenAI, those involved in lawsuits against the company, as well as unaffiliated publishers that either allowed or blocked ChatGPT’s search crawler” — the study found that ChatGPT search was correct in its title, publisher, date, and source link less than a quarter of the time.

(Source: CJR)

“We anticipated that ChatGPT might struggle to answer some queries accurately, given that forty of the two hundred quotes were sourced from publishers who had blocked its search crawler,” writes Klaudia Jaźwińska and Aisvarya Chandrasekar. “However, ChatGPT rarely gave any indication of its inability to produce an answer. Eager to please, the chatbot would sooner conjure a response out of thin air than admit it could not access an answer. In total, ChatGPT returned partially or entirely incorrect responses on a hundred and fifty-three occasions.”

The study said the inaccuracies took place “regardless of degree of affiliation with OpenAI,” pointing out that these issues of trustworthiness, brand safety, and work-recognition are likely not limited to publisher content.

“The flaws and inconsistencies in the way publisher content is currently accessed and represented in ChatGPT Search seem counterproductive to OpenAI’s stated goals and affect newsrooms regardless of how and if they chose to engage with OpenAI.”