Here’s a look at how some creators are leveraging generative AI to create video, audio and written content — and whether or not that’s a turn-off for advertisers.

AI-generated podcasts

Media and investment firm FlightStory — producer of popular podcasts such as Steven Bartlett’s “Diary of a CEO” — is market-testing seven AI-generated shows, including Bartlett’s “100 CEOs” interview series, on Spotify. These shows are fully scripted, produced and edited by AI, and hosted by Bartlett’s AI-cloned voice, although they are based on concepts and prompts written by Bartlett and his production team.

FlightStory is also experimenting with AI-generated animations for the visual component of “100 CEOs,” with plans to roll these out within the next three weeks, per FlightStory CRO Christiana Brenton. A human reviews the podcast at every stage of the production process to ensure that it feels authentic to Bartlett’s fans.

FlightStory CEO Georgie Holt said the company is currently amid a 60-day push to increase its use of AI, with internal teams competing to “disrupt themselves” by coming up with critical challenges and implementing AI to solve them. The goal is to make the company’s work more efficient and cost-effective by determining where human attention is no longer needed.

“I’m trying to kill myself in 60 days, essentially,” Holt said. “Can I remove myself entirely from the organization, and could it run without me, is what I’m trying to test.”

AI on YouTube

YouTubers such as Bennett “Money Mind” Santora are experimenting with combining different tools to create fully AI-generated videos. Santora uses ElevenLabs to clone his voice, HeyGen to produce AI video avatars and Poppy AI to write scripts, mixing the outputs of all three to make videos. He’s already uploaded hundreds of his past video scripts to Poppy AI to build a unique model around his personal tone and vocabulary.

Santora said that he was not overly concerned about his fans’ reaction to AI content because he believes they care more about the quality of the information he shares, rather than the specific tools used to create his videos. He plans to create a dedicated YouTube channel to share his AI creation journey with his fans.

“I’ll probably still need a human to go through the script and make sure that everything makes sense,” Santora said. “And also probably go through the ElevenLabs portion, because sometimes, the voice is weird in some spots, and I might just need him to make a couple of re-generations for those parts.”

Ultimately, Santora hopes to use AI tools to dramatically increase his video output, increase his share of search traffic and generate more revenue.

AI newsletters

Lauren DeVane regularly uses AI tools to write posts for her Substack blog, editing outputs from different models to produce blog posts about AI tech development and current events. She estimated that 70 percent of her Substack articles are AI-generated, and that all of the AI output was directly influenced by her personal ideas and context.

“I’m trying to see how I can use this to write stuff that really resonates. The more context you give it about who your audience is, the better it’s able to connect with them,” DeVane said. “I also know my audience well enough that I’m that taste factor — I’m that actual human that goes and looks at it and is like, ‘this hits, or this doesn’t hit.’”

Doing it right

Across podcasts, videos and Substacks, creators leveraging AI are putting effort into being transparent with their fans about it. Many people remain wary of the encroachment of AI into their daily lives. A 2023 report by creator marketing data platform IZEA found that 86 percent of consumers believe creators should disclose when they use AI to make content.

“The question of ethics and AI is obviously very multifaceted; I think everybody who’s not a writer uses ChatGPT, and it’s kind of their dirty little secret,” said Allison Harbin, a responsible AI expert who serves as an AI analyst for professional services advisory CBIZ. “From an ethical standpoint, I would say disclosing it is the right thing to do.”

In spite of the rise of AI-generated creator content, Harbin said that she anticipates fans will get better at discerning between real and AI-generated video and audio. As a result, she does not view AI-generated content as an existential threat for human-made content.

“Human-generated content will most likely win every single time, because of the tells, and because of the fact that a large language model is basically a parrot; it’s just parroting back information that it’s been trained on, based on the prompt that you’re giving it,” Harbin said. “So, I think that there are real creative limits to how good AI-generated content will be.”

The advertiser perspective

Otherwise, that could raise awkward conversations around creators’ fees. “They want to make sure the person they hired is the one actually making the content,” he said.

While some creators may be excited by the potential to scale their content output faster, the risk is that platforms will become flooded with poor-quality content if left unchecked. Harbin, who previously worked as a prompt engineer on Google’s AI search engine functionality, was unconcerned by this risk, anticipating that Google would eventually adjust its algorithm to accommodate it.

“Google is in the AI game as well, and they will tailor their search algorithms eventually to reflect that,” she said.

However, Whitt said that his clients were not perturbed by the use of AI-generated content to fulfill programmatic media buys, which are motivated more by the pure metrics than by the authenticity of a creator’s content.

“They don’t care if we use AI to fill a podcast ad slot somewhere across a network, or to crank out 100 versions of a banner,” he said. “But when it’s a person — especially a person with a face or a voice — then yeah, it starts to matter to them more.”

https://digiday.com/?p=578042