One of the most interesting and controversial trends in recent years has been the dramatic rise in the capabilities and proliferation of AI in various media – video, photo and text. AI is being integrated into our daily lives at every level – built into software, cars, digital assistants, social media platforms – you name it, it’s there.

But like anything revolutionary, this technology has drawbacks – and they’re big ones. For most of us, it’s like a black box. You give it an input and it spits out a response or an image. The quality of that response or output can vary dramatically depending on how well you give the AI ​​the first input, what kind of input you gave, what particular AI platform you chose, and how it was trained. Usually, the output is at least passable. I like to describe what you get from the AI ​​as being similar to the work of an experienced intern – you definitely need to check that, because the results can vary.

The problem is that all this information has to come from somewhere, be interpreted by various AI models and synthesized into a coherent result. The first problem: garbage in, garbage out. If the source information is wrong, so is the answer or output you get.

The second problem: There’s a lot of copyrighted material floating around on the internet, and when AI scrapes that source data—scraping in this case means automatically retrieving information from a website—the result may fall under the definition of “plagiarism.” That’s not a problem if you’re looking for the best pizza places in Old Forge, but it’s certainly a problem if you want to write a blog or create digital artwork with it.

Perhaps most concerning, sometimes even the developers of these AI platforms cannot clearly explain why they give the answers they give. AI is capable of teaching and training itself to a certain extent, so you end up with what you might call “emergent properties” – unexpected results or outcomes that arise due to complexity. No, I don’t think AI is about to take over the world. But I do think that the nature of the technology requires some guidelines, in particular that it should be able to explain why you get the answers you get and where they come from.

I tend to be cautiously agnostic about this topic – it’s a great assistant that saves a lot of time, and it’s great for quickly finding solutions or information that requires more detail than a typical Google search.

The danger is that while you can currently often – but not always – tell that you’re looking at AI-generated content or images, technology is getting better at producing human-like results. The risk is that someone could theoretically just type “give me a picture of person X doing questionable thing Y” and it would spit out exactly what you want, and you could present it as a real photo or video. AI detectors already exist, but they’re far from foolproof.

As with any technology, you can’t put the genie back in the bottle, and it’s already reached a point where you can circumvent the regulations on its use. The moral – and pretty much the only thing we can do in an increasingly ambiguous world – is to think critically about what we see or read.