Artificial Intelligence versus Maya Angelou: Experimental evidence that people cannot differentiate AI-generated from human-written poetry

TL;DR

This study shows that people often cannot distinguish AI-generated poetry from human-written poetry, especially when the AI's best work is selected, highlighting the advanced human-like quality of modern NLG algorithms.

Contribution

It provides empirical evidence that state-of-the-art NLG models like GPT-2 can produce poetry indistinguishable from human work in certain conditions, and introduces a novel incentivized testing methodology.

Findings

Participants failed to detect AI poetry when the best samples were selected.

People showed slight aversion to AI-generated poetry regardless of transparency.

AI-generated poetry was often judged as comparable to human poetry in quality.

Abstract

The release of openly available, robust natural language generation algorithms (NLG) has spurred much public attention and debate. One reason lies in the algorithms' purported ability to generate human-like text across various domains. Empirical evidence using incentivized tasks to assess whether people (a) can distinguish and (b) prefer algorithm-generated versus human-written text is lacking. We conducted two experiments assessing behavioral reactions to the state-of-the-art Natural Language Generation algorithm GPT-2 (Ntotal = 830). Using the identical starting lines of human poems, GPT-2 produced samples of poems. From these samples, either a random poem was chosen (Human-out-of-the-loop) or the best one was selected (Human-in-the-loop) and in turn matched with a human-written poem. In a new incentivized version of the Turing Test, participants failed to reliably detect the algorithmically-generated poems in the Human-in-the-loop treatment, yet succeeded in the Human-out-of-the-loop treatment. Further, people reveal a slight aversion to algorithm-generated poetry, independent on whether participants were informed about the algorithmic origin of the poem (Transparency) or not (Opacity). We discuss what these results convey about the performance of NLG algorithms to produce human-like text and propose methodologies to study such learning algorithms in human-agent experimental settings.