AIReal or Fake Text? We Can Learn to Spot the Difference

By Devorah Fischler

Published 8 May 2023

The most recent generation of chatbots has may have increased anxiety in areas such the creative economy and education, but the truth is that the effects of large-scale language models such as ChatGPT will touch virtually every corner of our lives.

The most recent generation of chatbots has surfaced longstanding concerns about the growing sophistication and accessibility of artificial intelligence.

Fears about the integrity of the job market — from the creative economy to the managerial class — have spread to the classroom as educators rethink learning in the wake of ChatGPT.

Yet while apprehensions about employment and schools dominate headlines, the truth is that the effects of large-scale language models such as ChatGPT will touch virtually every corner of our lives. These new tools raise society-wide concerns about artificial intelligence’s role in reinforcing social biases, committing fraud and identity theft, generating fake news, spreading misinformation and more.

A team of researchers at the University of Pennsylvania School of Engineering and Applied Science is seeking to empower tech users to mitigate these risks. In a peer-reviewed paper presented at the February 2023 meeting of the Association for the Advancement of Artificial Intelligence, the authors demonstrate that people can learn to spot the difference between machine-generated and human-written text.

Before you choose a recipe, share an article, or provide your credit card details, it’s important to know there are steps you can take to discern the reliability of your source.

The study, led by Chris Callison-Burch, Associate Professor in the Department of Computer and Information Science (CIS), along with Liam Dugan and Daphne Ippolito, Ph.D. students in CIS, provides evidence that AI-generated text is detectable.

“We’ve shown that people can train themselves to recognize machine-generated texts,” says Callison-Burch. “People start with a certain set of assumptions about what sort of errors a machine would make, but these assumptions aren’t necessarily correct. Over time, given enough examples and explicit instruction, we can learn to pick up on the types of errors that machines are currently making.”

AI today is surprisingly good at producing very fluent, very grammatical text,” adds Dugan. “But it does make mistakes. We prove that machines make distinctive types of errors — common-sense errors, relevance errors, reasoning errors and logical errors, for example — that we can learn how to spot.”

The study uses data collected using Real or Fake Text?, an original web-based training game.

This training game is notable for transforming the standard experimental method for detection studies into a more accurate recreation of how people use AI to generate text.