Blurring The Lines: AI-Generated Content vs. Human Content

Welcome to this week’s Deep-Fried Dive with Fry Guy! In these long-form articles, Fry Guy conducts in-depth analyses of cutting-edge artificial intelligence (AI) developments and developers. Today, Fry Guy dives into the discussion about AI-generated content vs. human-made content. We hope you enjoy!

*Notice: We do not receive any monetary compensation from the people and projects we feature in the Sunday Deep-Fried Dives with Fry Guy. We explore these projects and developers solely to showcase interesting and cutting-edge AI developments and uses.*

AI-generated content is going crazy. Everywhere you turn, there is another AI-generated image, video, or song going viral. In fact, by the year 2026, experts project that 90% of all of social media content could be AI-generated. Efforts to control it have been futile, and the AI tools being used to create such content are only getting better. But what exactly is AI-generated content? And is it really that much different than what has already existed for years? A deeper reflection upon these issues may reveal that what we think of as “AI content” might not be all that different than what we commonly take to be human-made. Let’s think about this together!

THE AI-GENERATED CONTENT WAVE

AI has displayed a unique ability to learn from human data to produce unseen content, including images, videos, text, and music. This use has extended to industries of all kinds and has gained a widespread adoption. But why has it gotten so popular?

One reason AI content has gotten so popular is because creation of such content is easy. You don’t have to be a computer programmer or musical genius to go to Udio.com and type in a prompt for a new song. In seconds, with the help of AI, anyone can create a new hit track. Want to write an essay for school? Type in a simple prompt. Want to create a new video? AI can do it for you! Content creation has never been so easy for people, and because the average Joe can use widespread tools like this for free, it makes such content creation more accessible to a wider variety of people.

Another reason AI content has grown in quantity is because it is trendy. Everyone loves jumping on a new trend and being a part of big events. People suddenly turn into football fans when the Super Bowl rolls around, and everyone becomes a political expert during election seasons. There is something uniquely human about jumping onto the latest trends and wanting to engage with them. People are fascinated by AI content because it is marketed as new and shiny, even though most of it isn’t even that high quality. In other words, one of the reasons AI content continues to grow in popularity is because it’s the cool thing to do! Once its novelty wears off, most people will probably resort back to a desire for genuine human content again. But for now, if you aren’t using AI, you are not part of the cultural wave.

STOP CALLING EVERYTHING “AI-GENERATED”

What many people fail to realize is that much of the “AI-generated content” on the internet is not actually “AI-generated.” Generative AI is a type of AI that can create “new” content, such as text, images, videos, and audio. It does this by learning from human data. However, this type of content is only a fraction of what most people call “AI-generated.” This has caused massive confusion for people like cousin Bob and grandma Sally, as they try to sort out this emerging technology.

To illustrate this point, take the example of edited photos. There are hundreds of AI photo editors online, and many of them are free to use. Most of these editors can take objects out of photos, change the facial expressions of people in photos, enhance the photos to make them more aesthetic, and more. AI can make these functions as easy as circling an unwanted object or typing in a prompt for how you would like the picture edited. Photos edited in this fashion are typically labeled as “AI-generated” or flagged as “AI content.” However, this isn’t quite accurate.

The keen observer will notice something peculiar about these types of AI editors: they don’t actually produce AI content. In fact, these editors are not much different than traditional photoshop, which has been around for almost 40 years! Of course, AI makes many of these processes easier, but the result is not much different than that of editing tools we have used for years. For instance, 90% of young women edit their photos before posting them on social media. This oftentimes includes the use of general photo editing apps, built-in social media filters, or simple iPhone photo editing capabilities. However, none of this content is flagged or raises an alarm for the public. But as soon as AI is used in the editing process, the content magically becomes “AI-generated.”

To think of this from another perspective, consider AI’s use in music. Dub-step is entirely computer generated, and artists like T-Pain have made their living off auto-tune. But now, because the word “AI” is used to label a sort of music creation and editing process, it is being viewed in a bad light. In fact, the Grammy’s banned songs from winning awards if they use too much AI. Meanwhile, T-Pain has won two.

These discussions blur the lines between what we consider “AI-generated content” and “human-made content.” If auto-tune and filtered images on Instagram are regarded as “human-made,” then what is so different about AI content that separates it? Some people might say, “Well, maybe AI editing is okay, but the creation of something new entirely is deceptive.” But is not the advanced editing done for decades on photos, music, and videos equally deceptive to the public as AI-generated content? Isn't it deceptive to post a selfie on Tinder that misrepresents one’s appearance? Why is one form of deception okay, but the other is not?

To further this point, consider how a human learns compared to how a generative AI model learns. Like AI, humans learn from a variety of inputs—the inputs just look different—and use those inputs to produce outputs. For example, the human student learns from their professors in classes, and then writes an essay. Is the human not then just reflecting various inputs? Is the “human-made” output any less reliable given that it is only a reflection of what they have learned? Maybe AI content isn’t so different. AI is trained on human-made content, so one could argue that AI never produces anything “new,” it merely reflects what it has been taught by humans. At the end of the day, maybe it’s not AI’s fault when it goes awry—maybe it’s the fault of the humans who produced the content it was trained on. And if that’s the problem, we shouldn’t trust human-made content any more than we should trust AI content.

WHY DO PEOPLE HATE AI CONTENT?

Whether or not you agree that AI content is not too much different from human content, there is no doubting that AI gets especially sensitive treatment when it comes to content creation. There are multiple reasons that people have grown overly sensitive to AI content, but two in particular drive the public hysteria. One of these is the association of “AI” and “fake.” In reference to content, these two words are used almost interchangeably. When someone says, “That content is AI,” what they typically mean is something like, “That content is fake.” It’s easy to see why this might be the stigma surrounding AI. When people think about AI-generated content, they often think about fake articles, fabricated images of celebrities, or videos meant to deceive the public. AI has sometimes been used to deceive or has produced absurd outputs, such as telling people to use gasoline in place of olive oil to cook their spaghetti. This can lead to a mistrust of AI models as something that produces only fake content.

Even though AI has the potential to produce “fake” content, the equivalence of “AI” and “fake” is overblown. Generally, AI models produce content that is accurate and of high quality. In fact, it is outsmarting many humans in mathematics and has shown the ability to pass MBA exams. So yes, AI can produce “fake” outputs, but AI content is not equivalent to fake content, as many think.

A second reason people are sensitive to AI content is because of copyright concerns. People are worried that AI-generated music, for instance, steals from the hard work of humans who created the music it learns from. The same is said for AI art, video, voice, and more. If people believe this content is—in many cases—stolen from humans without proper compensation or attribution, it comes as no surprise why they view it negatively.

RETHINKING HUMAN-MADE CONTENT

AI-generated content is booming, and its negative connotation means many people view this is a bad thing. But before you start labeling all AI content as “fake” or “bad,” remember that much of AI is generally reliable and reflects a type of editing done by humans for decades. And as the lines blur between AI-generated content and content that is “authentically human,” you might find it worthwhile to reflect on what the difference is, and if there even is one at all.