Last week, as I scrolled through Instagram, I came across a profile that caught my eye. At first glance, like every other AI profile, it had hundreds of generated images. This isn't something shocking in this day and age. The weird part was that the profile did belong to an actual person. From what I understand, they were using AI to edit or generate photos of themselves. This got me thinking about the things to come.
Even if you aren't using AI to generate content, you are likely using it indirectly in some way. So, more and more, we become reliant on AI. Worse, anyone claiming to reliably differentiate whether something is AI-generated is lying to you. Most automated detectors are brittle and accurate only under narrow conditions. The amount of generated content will only increase, so this problem will only worsen. Now, solutions like cryptographic provenance exist, but they are not widespread. Many of these solutions will become more widespread in the next few years. Until then, we assume there are no truths.