I don’t think AI is going to make disinformation worse.

Will AI Really Worsen the Disinformation Problem? A Critical Perspective

In recent discussions about the role of artificial intelligence in our digital landscape, a common concern revolves around AI’s potential to amplify disinformation. Many worry that AI-generated content, flooding social media platforms, will make it increasingly difficult to discern truth from falsehood. While this concern is understandable, I believe the situation might not be as dire as it appears.

The Rise of AI-Generated Content

It’s true that AI has the capacity to produce vast amounts of “junk” content, often indistinguishable from human-created posts. When we observe social media feeds as a whole, the presence of AI-produced material has grown significantly. Naturally, this leads to fears that disinformation — deliberately false or misleading information — will multiply correspondingly.

However, a closer look suggests that the impact may not be as profound as anticipated. Consider the typical user behavior: if I spend a certain amount of time scrolling through TikTok or similar platforms, I might encounter between 100 and 150 short videos. Whether these videos are created by humans or generated by AI, the sheer volume remains roughly the same. Injecting AI-generated content into the mix doesn’t necessarily increase the total number of videos I view.

The Role of Human-Generated Disinformation

It’s important to recognize that humans have been producing an enormous amount of disinformation long before AI’s rise. Our digital environment has already been saturated with false narratives, propaganda, and biased content at an unprecedented scale. Adding another petabyte of artificial content doesn’t exponentially increase my exposure—my attention span and content consumption habits remain constant.

My interests and preferences guide what I watch, regardless of whether the content is real or AI-generated. I tend to spend my time on a mix of entertainment, such as cat videos, humorous fails, or political snippets. The distribution remains similar; only the source changes. In essence, my exposure to disinformation over the last five years wouldn’t have significantly altered if AI-generated content had been present earlier.

The Subtle Masks of Disinformation

One of the more insidious aspects of AI-driven content is how it can subtly facilitate misinformation. For example, edited clips of celebrities or politicians saying things they never uttered can appear convincing, especially when presented in formats that blend seamlessly with legitimate content. These edited snippets might not look immediately deceptive—they are more nuanced than blatant lies. This can make disinformation harder to spot, but it also depends on how many such clips are produced and circulated