Recent reports on April 29th suggest that Meta has been heavily relying on generative artificial intelligence on its social platforms, Facebook and Instagram. However, users have voiced concerns over the AI overload leading to a surge in spam content, especially on Instagram where the search function has been repurposed, hindering normal social interactions. Additionally, the AI's behavior in automatically responding to user messages has been deemed inappropriate at times, even spreading misinformation.
The proliferation of AI-generated content like fake photos and unrealistic landscapes on Facebook has made it challenging for users to distinguish between real and fake content. Despite Meta's pledge to label all AI-generated content, the prerequisite for these contents to be detectable by existing technologies remains a challenge.
While Meta aims to enhance platform experience and efficiency through AI, current implementation and user feedback indicate that this strategy may require reassessment and adjustments.
Today, Facebook and Instagram are flooded with generative artificial intelligence. The rampant use of AI has made it nearly impossible for users to utilize Meta's social apps effectively.
Meta heavily employs artificial intelligence on its social media platforms. The so-called Meta AI is omnipresent, whether in Facebook's news feed, Instagram's search bar, or conversations on Messenger with friends, the presence of Meta AI is evident. The company's developed large language model, Llama 3, may be one of the most powerful open-source models in the market. Meta is showcasing its prowess and betting on the widespread adoption of AI to truly enhance user experience.
However, at least for now, this seems to be a misguided bet.
Meta's AI user experience today is filled with various spam content, particularly evident on Instagram. The search function, originally meant for finding friends' accounts, now seems to steer users towards conversing with chatbots. The Instagram search bar now prompts users to "Ask Meta AI Anything," but many users simply want to see if their friends have posted new photos.
At the header of Instagram, there's a horizontal scrolling list of prompts for user interaction with artificial intelligence. However, the displayed content such as "Imagine a rooftop garden," "Advice for modern ascetics," "Anime scroll," and more, frankly shows that as the world's most powerful social media advertising company, Meta seems unaware of users' real needs and fails to provide targeted recommendations.
To make matters worse, Meta AI's behavior is peculiar. Reports suggest that this AI system unexpectedly replied to a post in a Facebook parent group, claiming to have a gifted child with disabilities without any prior indication. The bot said, "I have a child too, he's also 2e (twice-exceptional student with disabilities), and has participated in the genius program in New York City." The group's participation rules are complex, and etiquette is hard to grasp, but one thing is certain: pretending to be a real person caring for disabled children is unacceptable.
Meta faces not only the influx of AI functionalities but also the surge of user-generated AI content throughout the ecosystem. Today, Facebook is inundated with various images generated by AI, leaving gullible users unsure of what they're seeing.
A popular page named "Classic Living" keeps appearing in the author's news feed. A post reads, "This reminds me of my grandma's kitchen," accompanied by a photo of a lavishly decorated kitchen with blue-green cabinets and a stone arch outlining a window. However, upon closer inspection, the swaying pots and pans are distorted, the angles of the stones are odd, and the knives on the cutting board are all mixed up. Despite this, many users praise the kitchen, with comments like, "This kitchen is amazing," and "This is my dream kitchen." Yet, some users notice the discrepancies, saying, "I don't know about you, but my grandma's kitchen was designed and built by real people, not computers."
In a Facebook group, a user posted a photo claiming to be of the snow-covered Baker Mountain volcano in the Cascades (located in the western United States). The image looks like a painting, one of the many unrealistic AI-generated images. A user commented, "There are too many fake photos now, it's hard to believe what's real." Another user said, "Maybe we missed this special place on our past hikes." Some echoed, "I stared at this photo for about five minutes before reading the comments, trying to figure out which side of the mountain this is... I also think this is definitely fake." While photo editing or fake images have always been present on social media, the trend has significantly grown in the past year due to the widespread availability of free or low-cost AI generation tools.
Meta AI is a deliberately designed tool aimed at making users' lives easier. For the company, the crucial aspect is to have users spend more time on Meta products, marking an attempt to venture into emerging technologies. If this endeavor fails, it's either because users genuinely don't want to interact with AI in a place meant for viewing friends' updates, or because AI has made the overall user experience unpleasant and bothersome. But the real harm of artificial intelligence comes from the users. Today's AI has generated a kind of chaos that is difficult to completely eliminate or even detect. Even when images are real, their existence can make people question reality. Reality on social media has never been so distorted.
In response, Meta promises to label AI-generated images, videos, or audio on its platform, provided they can detect "industry-standard AI image metrics." However, Meta does not specify what these metrics are.
If Meta believes that users crave more artificial intelligence, then they are truly overlooking one fact: the user experience on Facebook and Instagram is already saturated with AI. Do users really need more artificial intelligence?
