In a 22-minute video circulating on social media, a range of impeccably presented news anchors take center stage, delivering the day’s news. However, there’s a unique twist – none of these anchors are real. The footage seamlessly aligns with content seen on numerous global news channels; they are not real humans; instead, they are the product of artificial intelligence (AI) technology.
The video arises from Channel 1, a Los Angeles-based start-up founded by entrepreneurs Adam Mosam and Scott Zabielski. The duo will introduce AI-generated news content on a streaming TV channel later this year. Mosam explains that there’s a compelling opportunity to enhance the user experience of news by leveraging AI to customize content for individuals.
In addition to personalization, AI technology showcases its versatility by assisting in language translation for scripts and interviews. Channel 1 demonstrated these capabilities in a December promotional video, highlighting AI’s potential to make news more accessible and tailored to various audiences.
Channel 1 illustrates the growing trend of AI-powered news presenters worldwide. In Kuwait, an AI persona named Fedha delivered headlines for Kuwait News. In May 2023, Hermes took the news anchor role for the Greek state broadcaster ERT. South Korean broadcaster SBS enlisted Zae-In, an AI-generated deepfake, to handle news presenting duties for five months this year. Similar AI-generated news presenters have emerged in India and Taiwan, marking a global transformation in delivering news, with AI taking center stage.
Can Viewers Trust AI-Delivered News Over Human Anchors?
We’re not doing this because we think a robot performs better than a human – Adam Mosam.
Trust in human news presenters has hit a historic low, as revealed by a public opinion polling firm, Ipsos Survey. In the UK, only 42% of people now trust TV newsreaders, marking a significant 16-percentage-point decline in just one year. This suspicion regarding news presenters as impartial purveyors of truth reflects a contemporary trend, with an increasing number of individuals opting to receive their news from individual creators or influencers.
These social media influencers produce a unique connection with their audience, known as parasocial influence. Stamped in the 1950s by academics at the University of Chicago, parasocial relationships were initially described as the belief held by viewers of nightly news programs that the anchor behind the desk was speaking directly to them through the camera. News presenters transformed from sheer conveyors of information to trusted friends welcomed into viewers’ living rooms night after night.
The concept of journalists not physically reading the news is somewhat unknown, even though the idea of them being computer-generated is. Nic Newman, a Senior Research Associate at the Reuters Institute for the Study of Journalism at the University of Oxford and a former BBC editor, points out that in the early days of journalism, actors were often the ones reading the news, and people generally accepted that.
Newman believes the current trial of AI-generated news anchors might succeed, with certain limitations. He anticipates its effectiveness for brief news bulletins. However, he suspects viewers readily embrace a parasocial relationship with an AI anchor for delivering comprehensive news programs. Newman emphasizes that the human touch and connection will likely remain crucial, especially for news.
Christine H Tran (University of Toronto) shares a similar uncertainty about the potential impact of AI personalities. They question whether AI broadcasts labeled as ‘AI content,’ with viewers aware of the absence of a personal life beyond the screen, can enable the same parasocial connection. The outcome may hinge on whether platforms hosting AI presenters adopt content labeling practices, similar to considerations made by platforms like Instagram moving towards such transparency.
For Channel 1 and NewsGPT, self-proclaimed as the world’s first AI-generated news channel, another question looms: Can the human element be entirely removed from the equation?
Channel 1 boasts nearly a dozen staff members dedicated to reviewing AI-generated scripts and curating stories for coverage. According to Mosan, the company follows a meticulous 13-step process for each story to address potential issues tied to generative AI before airing. It includes combating hallucination, where AI tools create fictitious content, a practice incompatible with journalistic standards. The company plans to hire an editor-in-chief early in the upcoming year.
AI’s ability to identify and report on newsworthy events is a challenge acknowledged by Mosan and Newman. The test episode from Channel 1 heavily relied on stories and footage from human journalists. Newman points out that AI needs these human sources to avoid significant limitations. “If that raw material is not there, then the AI has absolutely nothing to work on,” he observes.
While Mosan believes AI can handle certain aspects of the reporting process, he acknowledges its limitations. He notes that gathering intelligence and conducting person-to-person interviews effectively are tasks that AI may need to replicate. However, he sees potential in using AI for tasks such as flying drones and analyzing visual data. Channel 1 only plans to rely on AI for newsgathering, with human involvement.
