AI scripts on social media: the video that looked good and became a joke

A 60-second video script requested from a chatbot arrives in seconds, with an opening question, three lessons and a catchy line at the end. Recorded in the hallway and posted the same day, it runs the risk of getting few views, likes from the team itself and no comments from customers, plus the question nobody wants to hear: whether ChatGPT wrote it.

We wrote this 12 article series about that moment. We look at videos and posts by leaders, managers and brands on Instagram, LinkedIn, YouTube and Facebook when the script is born from a generic request to an artificial intelligence tool and goes live without anyone putting an idea of their own into it. You may have already approved a text like that. It looks good at approval time and reaches the feed looking like hundreds of other videos your audience has already seen.

How much AI text already circulates on professional social networks

Measuring the presence of AI on social networks is hard, and the available numbers come from companies that sell detectors. Even with this caveat, they point to a clear trend. In 2024, Originality.ai analyzed 8,795 public LinkedIn posts with more than 100 words, published between January 2018 and October 2024. According to Fast Company, the study estimated that more than 54% of long posts in English were probably generated or edited by AI, with a 189% jump from January to February 2023. In July 2026, the same company repeated the exercise with 5,000 posts and classified 81.2% of them as "likely AI".

These percentages come from a classifier, which estimates probability without proving authorship, and Originality.ai sells detectors. In our view, the numbers work as a thermometer. They indicate that when you open a professional network, you live every day with a lot of text produced the same way.

When the generic AI script becomes a joke

The most visible risk is public embarrassment. In July 2026, according to Daily Hive, Bill Oliver, a provincial legislator in New Brunswick, Canada, read in the chamber a text that included ChatGPT's own comment offering "a more natural and fluid version" of the passage, one that sounded like a legislative speech. The video spread on Reddit. The fifth article of the series brings together this and other cases documented by the press.

There is also silent damage. A video with a generic script rarely makes the news. It goes through the feed without anyone stopping, commenting or sharing, and the conversation about it happens far from your brand, in groups and private messages.

What the platforms say about generic and repetitive content

Here we ask for precision, because the topic attracts exaggeration. The official rules we checked for this series target content that is unoriginal, recycled, repetitive, mass produced or low quality. None of the official Meta pages we consulted states that a text written with AI, in and of itself, has its reach reduced.

What these pages do say already carries a lot of weight if you post often. The Facebook for Creators blog, in July 2025, stated that "authentic stories resonate with audiences and are more likely to perform well". Instagram, in April 2026, defined original content as work "that you created entirely or that reflects your unique perspective". YouTube, in July 2025, renamed one of its monetization policies to "inauthentic content" and included among its examples AI generated content "with generic or unoriginal templates", which gives the impression of mass production.

LinkedIn went further. In July 2026, the network created an option for users to flag a post as "Seems like AI slop", an English expression for AI content with no substance. In August, according to Engadget, product chief Hari Srinivasan said members were seeing 40% fewer views of what the platform classifies as slop, compared with a few weeks earlier.

Distribution and connection with your community

A video on social media depends on two things that go hand in hand. The first is distribution: the platform decides whether to show the post to people who do not yet follow your profile. The second is connection: the viewer stops, feels something, comments, saves and comes back the following week. A generic script tends to weaken both. It resembles what the platforms describe as repetitive and lacking a perspective of its own, and it asks little attention of the viewer, because it offers nothing the person has not already seen.

Think about who makes up your community on social media: customers, employees, job candidates, suppliers and competitors. These people know your voice, the company's stories and the feel of your city. When the video brings lines that could be on any profile, they may notice the distance between you and the text you read to the camera.

The counterpoints we take seriously in this series

Some data call for caution, and we make a point of showing them. In 2024, Buffer analyzed 1.2 million posts and found higher median engagement in those made with the help of the tool's AI assistant: 5.87%, against 4.82% for posts without AI. Buffer itself pointed out the bias in the result, because the most active users are the ones who use the assistant most, and the gain would come from volume and posting consistency. In 2023, a study released by MIT Sloan showed that participants who did not know the origin preferred texts by GPT-4.

These results indicate that the tool writes well and helps people who post regularly. None of these measurements examined the situation we deal with here: a script with no story, no opinion of its own and nothing that only that leader could say.

What the next articles in the series will show

The second article takes a close look at Meta's guidelines on original content, demoted content and engagement bait. The third explains why platforms deliver less of what is generic and repetitive. The fourth gathers the ChatGPT phrases the audience already recognizes in videos and captions. The fifth brings the documented cases of AI texts that became jokes, and the sixth details LinkedIn's offensive against AI slop.

In the second half, we follow what the head of Instagram has been saying about authenticity, the studies on trust when followers suspect AI was involved, the structure of a Reels script with a voice of its own, the responsible use of AI in the process and a review checklist before posting. The last article talks with you about when it is worth having a scriptwriter for social media.