Can You Trust AI News? The Risks of Automated Journalism

Technology Policy Brief #105 | By: Mindy Spatt | December 26, 2023

Photo taken from: www.rollingstone.com

__________________________________

THE RANKS OF TRADITIONAL NEWS MEDIA HAS SHRUNK in the Internet age as news became easily available for free or at low cost online. Now that artificially created news content can be had for the taking will it become a go-to source for media and cause newsrooms to disappear altogether?

Analysis

My first experience with ChatGPT was asking it a version of the question above. I received a very bland and reassuring answer with just a tad of an ominous warning tacked onto the end. “While AI may automate certain aspects of journalism, it’s unlikely to replace journalists entirely. AI can assist in tasks like data analysis and content generation, but human intuition, ethics, and the ability to tell compelling stories remain essential. Journalists may need to adapt their skills to work alongside AI tools….”

My second question: How accurate is news created by AI? “The accuracy of news generated by AI largely depends on the quality of the underlying algorithms and the training data used,” the chatbot explained. “While AI can efficiently summarize information, it may still face challenges in interpreting context, bias or understanding nuanced details.” I was impressed with this humble tone. The response also warned that “It’s crucial to rely on reputable sources for critical news.” But how do I know if ChatGPT is getting its information from a reputable source?

According to Newsguard, a media misinformation watchdog, websites masquerading as news sites are finding it easier and easier to push out AI generated content to growing audiences. In a press release announcing its new AI Tracking Center, NewsGuard co-CEO Steven Brill said “…AI-generated ‘news’ websites are akin to the misinformation-spreading Macedonian content farms from a few years ago, with the distinction that these new sites cost even less to produce and can become even more prolific with the enhancements of artificial intelligence.”

At launch earlier this year the tracker featured 150 sites deemed unreliable. At latest count (Dec. 18), that number was 614, and included sites in Arabic, Chinese, Czech, Dutch, English, French, German, Indonesian, Italian, Korean, Portuguese, Spanish, Tagalog, Thai, and Turkish.

According to Newsguard, the offending websites often have generic names, such as iBusiness Day, Ireland Top News, and Daily Time Update, which obscure that they are created with “little to no human oversight” and publishing content “written largely or entirely by bots.”

In addition to being unsourced, the chatbot’s response to my question failed to reference or acknowledge industry concerns or trends. The Writer’s Guild is certainly worried; their new contract, won after a grueling 148 day strike, contains robust AI protections, including that studios will not be able to use generative AI to write or rewrite literary material and AI-generated content cannot be used as source material.

It is hard to gauge how deep AI’s inroads into newsrooms will be, but here too things are moving quickly. The New York Times, one of many news outlets that have blocked their content from ChatGPT, recently hired an editorial director of artificial-intelligence initiatives with the goal of learning “how we do and do not use generative AI.”

Axel Springer, the publisher of Business Insider and Politico, just announced a very high-profile deal in which ChatGPT will be allowed to summarize its content for users, who, using the free chatbot, will be able to access material otherwise blocked by a paywall with links to full articles.

Perhaps that deal was already in the works when Springer CEO Mathias Dopfner warned, in March 2023, that AI would be better at aggregating information than humans. The role of journalists, he said, would be to understand people’s ‘true motives.’ “Only those who create the best original content will survive,” Dopfner predicted. That sounds an awful lot like what the chatbot said, doesn’t it?

__________________________________

Engagement Resources

• Writers Guild Wins Protections Against Artificial Intelligence By: Tom Jones and Angela Fu, September 28, 2023
• Tracking AI-enabled Misinformation: 614 ‘Unreliable AI-Generated News’ Websites (and Counting), Plus the Top False Narratives Generated by Artificial Intelligence Tools

Get the latest updates from our reporters by subscribing to the U.S. Resist Democracy Weekly Newsletter, and please consider contributing to Keeping Democracy Alive by donating today! We depend on support from readers like you.