AI-generated text is overwhelming American institutions

Companies accepting written submissions from the internet are essentially running an all-you-can-eat buffet for AI-generated content, according to a recent series of reports and data.

The issue: Institutions across the country have reported struggles dealing with an overwhelming amount of AI-generated text submissions.

A few examples:

• Newspapers and academic journals are being flooded with AI-generated letters to the editor.
• Lawmakers struggle to discern between human- and AI-written emails from purported constituents.
• Scientific conferences are seeing a rise of AI-made ‘sloppy’ research papers, especially in subjects relating to AI (an ironic twist).
• Courthouses and lawyers report an increase in AI-generated filings and legal advice.
• Social media sites are inundated with AI posts from bots and spammers.

AI generated content is also being overwhelmingly produced in music, education, investigative journalism, and hiring.

Institutions are responding in a few ways. Some have slammed the door shut on submissions altogether, at least temporarily. Others are fighting fire with fire and using AI tools to screen writing submissions, moderate content, review job applications, grade assignments, and administer exams—though AI-detection software is notoriously unreliable.

Looking ahead…Analysts say the influx of AI-generated content—and the arms race to develop systems that can identify such content—is poised to continue for the foreseeable future as generative AI tools continue to improve.