How an AI-generated ‘expert’ sank into media deadlines

Ashley Abramson first came across Sophie Cress in a cold pitch to her work email. Cress was asking to be an expert source for any stories Abramson was working on as a freelance reporter. “I’ve got over 8 years of experience and qualifications in Psychology and Couples & Family Therapy, and I’m enthusiastic about exploring potential collaborations, especially in the areas of love, relationships, or LGBTQIA+ topics,” Cress wrote. 

She provided a list of links to articles where she’d supposedly been featured as an expert. Her email address, linked to a website reviewing sex toys, caught Abramson’s attention. Then, when Abramson insisted that she could only conduct interviews over phone or video call, Cress ghosted.

In a recent investigation for Allure, Abramson dug deeper into Cress’s background and alleged qualifications. Turns out (surprise!), she doesn’t actually exist and was created by the Latvia-based owner of sex-toy-review site, Sexual Alpha, to boost traffic and improve the site’s search rankings.

Dainis Graveris, owner of Sexual Alpha, did not respond to Abramson’s requests for comment but Abramson decided to investigate further. She started by searching for evidence of a “Sophie Cress” or similar names licensed in North Carolina or holding the degrees and certifications Cress claimed. She found none. Abramson also discovered that Cress’s headshot was a stock image, and the woman pictured was not named Sophie Cress.

Most journalists contacted by Cress simply took her at face value, allowing her operators to dupe outlets from the Metro to the Daily Mail. As Abramson writes, “Of course, anyone could always claim to be anyone, and AI programs make it easy to generate a chunk of text that seems, at least at first skim, like it was written by an expert in any field you can think of.”

This is a classic case of what is commonly known as internet slop: scammy, AI-generated content that’s becoming increasingly widespread online and beyond. Some studies have even found that people rated AI-generated content more favorably than content actually created by humans (or at least can’t tell the difference).  

This rising tide of slop only serves to further clog the internet, which is already drowning in misinformation. “While this is unlikely to be some sort of election-altering Russian disinformation campaign,” Abramson concludes, “I wouldn’t say it’s a sign of a particularly bright future.”

https://www.fastcompany.com/91266972/how-an-ai-generated-expert-sank-into-media-deadlines?partner=rss&utm_source=rss&utm_medium=feed&utm_campaign=rss+fastcompany&utm_content=rss

Creato 3mo | 24 gen 2025, 22:30:03


Accedi per aggiungere un commento

Altri post in questo gruppo

This startup wants to reprogram the mind of AI—and just got $50 million to do it

Anthropic, Menlo Ventures, and other AI industry players are

22 apr 2025, 11:10:02 | Fast company - tech
Duolingo will start teaching chess soon

Popular language learning app Duolingo is giving its bite-size lesson treatment to one of the oldest games

22 apr 2025, 08:40:04 | Fast company - tech
How tariffs could change the way Americans buy video games

In a time where tariff price hikes are invading seemingly every

21 apr 2025, 21:10:06 | Fast company - tech
The Ice Bucket Challenge is back, this time with a focus on mental health

Remember the viral “Ice Bucket Challenge” of 2014? Over a decade later, it’s back—

21 apr 2025, 21:10:04 | Fast company - tech
FTC sues Uber over misleading Uber One subscribers

The U.S. Federal Trade Commission on Monday sued Uber Technologies, accusing it of signing up some Uber One subscribers without their knowled

21 apr 2025, 21:10:04 | Fast company - tech
Big Tech enters earnings season bruised by Trump administration turmoil

As Big Tech kicks off its quarterly earnings season this week, the industry’s bellwether companies have been thrust into a cauldron

21 apr 2025, 18:40:11 | Fast company - tech