{"id":308,"date":"2021-02-24T16:38:10","date_gmt":"2021-02-24T16:38:10","guid":{"rendered":"https:\/\/connect.apsanet.org\/s42\/?p=308"},"modified":"2021-10-18T13:03:43","modified_gmt":"2021-10-18T13:03:43","slug":"all-the-news-thats-fit-to-fabricate-ai-generated-text-as-a-tool-of-media-misinformation","status":"publish","type":"post","link":"https:\/\/connect.apsanet.org\/s42\/2021\/02\/24\/all-the-news-thats-fit-to-fabricate-ai-generated-text-as-a-tool-of-media-misinformation\/","title":{"rendered":"All the News That\u2019s Fit to Fabricate: AI-Generated Text as a Tool of Media Misinformation"},"content":{"rendered":"\n

Sarah Kreps, Miles McCain, and Miles Brundage<\/p>\n\n\n\n

In October, 2020, Former CEO of Google, Eric Schmidt cautioned<\/a> that \u201cthe broad ability to do deepfakes, and fake texts and so forth, it\u2019s only going to get worse.\u201d    He was referring to images, video, and text that appear real but have been generated by machine learning algorithms. The comments were directed to the National Security Commission on Artificial Intelligence tasked with understanding how artificial intelligence can be used maliciously to manipulate public psychology\u2014referred to as malign AI and AI-powered information operations. <\/p>\n\n\n\n

Absent from the conversation, but a crucial first step in framing the problem, is to understand the technology: whether AI-generated text, for example, is even capable of passing as authentic text, let alone whether it can persuade people of a particular view, and whether disclaimers can successfully alert readers to the presence of AI-generated content. We answered these questions in a study<\/a> of GPT-2<\/a>, a natural language model that synthesizes original text based on previous words within the text.<\/p>\n\n\n\n

Since we had a particular interest in international politics, we first selected a story from the New York Times<\/em><\/a>\u2019 <\/em>story about a North Korean ship seizure as the input for generating synthetic text samples, using both one and two-sentence inputs to generate entire articles with the length, style, and substance of the original Times <\/em>piece.   See examples of the input text, the original article, and GPT2 output below<\/a>:<\/p>\n\n\n\n

\"\"<\/figure>\n\n\n\n

We then randomized the GPT-2 outputs\u2014using three different sized models\u2014and the New York Times<\/em> story and asked individuals about the perceived credibility of each story.  The text from the least powerful model was perceived as much less credible than the Times <\/em>piece, but the medium and large-size models were perceived as equally credible, and in one case more <\/em>credible, than the original Times <\/em>piece.<\/p>\n\n\n\n

Next, we elected to study a more controversial topic and investigate whether GPT-2 could be used to manipulate public attitudes in the spirit invoked by the National Security Commission on AI. Here we selected news outlets on the left, right, and center\u2014Huffington Post, Fox News, <\/em>and the Associated Press<\/em>\u2014as GPT-2\u2019s input.  In addition to randomizing the original story with the GPT-2 version, we included a treatment condition with a disclaimer on the synthetic text story, overall a 3×3 design that varied the ideological angle with authenticity of the text (original story, GPT-2, and GPT-2 + disclaimer).<\/p>\n\n\n\n

We again gauged perceived credibility of the news stories and then attitudes toward immigration to understand the potential political effects of exposure to both congenial and non-congenial content.<\/p>\n\n\n\n

Findings suggested the following:<\/p>\n\n\n\n