It seems you can’t even chat with friends on one of the world’s most popular social networks these days without hearing something about AI.
From deep fake videos to realistic-looking ‘news’ articles and even lifelike image filters that give you a celebrity makeover, there’s no shortage of AI tools gaining viral attention on social media.
So, what is Meta, the parent company of Facebook, doing to identify AI-generated content?
Additionally, how do you know if the content you see on Facebook and other social media sites is AI-generated?
Just as the name implies, AI-generated content is created by artificial intelligence such as ChatGPT, Google Gemini, or Claude.
AI tools are trained on vast amounts of content to analyze and generate human-like responses when asked questions.
Yet, for all of their training, AI text generators still have some common signs, including unnecessarily sophisticated language and a predictable writing cadence. The result? AI-generated content typically reads as wordy, even for simple questions.
At the same time, humans aren’t as good as they think they are at identifying common ChatGPT words. That’s where AI checkers are beneficial. They create transparency between AI-generated content and human writing, whether it’s posted on a blog or in a social media caption.
Text generation is only one part of generative AI! Artificial intelligence offers a range of capabilities for social media.
AI tools are often used in the name of good fun for social media. However, it’s important to note that AI can cause issues, so always review AI-generated posts before publishing and maintain transparency with your audience.
According to Meta, AI content is detected if it “contains industry-standard signals that it’s generated by AI…this includes content that is created or edited using third-party AI tools. It also includes content that is created using Meta’s AI tools...”
So, to answer the popular question, ‘Does Meta detect AI content?’ — yes, Meta detects AI content.
If Meta’s AI detection systems identify AI-generated content or AI-altered content, they’ll add an ‘AI info’ label to the post.
In 2024, Meta created a set of guidelines that describe how the company (and Facebook) approach content created with AI. The guidelines note that they were based on “feedback from the Oversight Board and our policy review process with public opinion surveys and expert consultations.”
As a brief snapshot of the guidelines, Meta wants to encourage transparency so users know when they see posts created with AI.
In July 2024, they introduced an update to their artificial intelligence label. The label, which says “AI info,” allows users to quickly see whether some form of AI tool was used to create the post.
Rather than impose restrictive rules, Meta is opting to add labels for transparency.
What happens when an image is created by a human and then slightly edited with AI (such as with a retouching app)? Originally, Meta labeled this type of AI content as “Made with AI.” However, with their July update, there is now an ‘AI info’ label for more information.
As they’ve done historically, Meta also enlists the help of their independent fact-checkers for original or AI-created content. According to Meta’s guidelines, “When fact-checkers rate content as False or Altered, we show it lower in Feed so fewer people see it, and add an overlay label with additional information. In addition, we reject an ad if it contains debunked content (sourced via Meta).”
Detecting Meta AI content (and AI content on social media generally) matters for several reasons.
Facebook has billions of users. In their 2023 fourth-quarter investor report, they noted that Facebook sees over 3 billion monthly active users.
So, trying to identify every instance of AI-generated content is a herculean task. The volume of content their systems need to process is one of the first initial challenges they face.
Then, there’s the possibility of false positives. When genuine content is flagged as AI, users can become frustrated that their legitimate posts were labeled as AI-created.
The reverse is also true, with AI detection sometimes slipping through the cracks and not being labeled as it should be. Given how sophisticated and human-like Meta AI content has become, it’s possible that some instances could fly under the radar.
Then, there can be delays if a review by a human moderator is required, which creates bottlenecks that may result in a drop in engagement.
Whether you’re publishing on Instagram or Facebook, there are a few best practices to follow.
Meta updates its content guidelines and policies regularly and notifies users of major changes. Staying informed about these changes can help you reduce the risk of having your content flagged for AI or otherwise.
Meta offers feedback channels. If an original post is flagged as AI, reach out to their feedback team to report the false positive and help improve overall accuracy.
Redirect users from the social media platforms to drive engagement on other channels, such as an email newsletter or landing page. This can also help you establish a content marketing funnel that drives sales or boosts traffic.
Meta has an extensive audience, which is a great way for content creators to connect with followers. However, it’s best to diversify instead of relying solely on one platform. Share your brand on several platforms to expand your reach.
Facebook, in particular, has several groups devoted to content creators, where members share tips, feedback, and guidance. They have the potential to be a great resource for navigating Meta AI content detection.
As developments in AI continue, Meta is responding by promoting transparency with ‘AI info’ labels for AI-generated or AI-altered content.
To follow best practices, always review AI-generated content before posting on social media.
Use the Originality AI Checker to detect AI-generated text. Then, before you hit publish, review AI content to ensure it's factual and maintains your brand voice.
