You Ask, I Answer: Detection of AI Content?

Xiaoli asks, “How can people determine whether some content is created by AI or human?”

In this video, I answer a question from a viewer who wants to know how to determine whether content has been created by AI or by a human. While there are tools available online that can detect language patterns of generic models, their quality is hit or miss, and they are tuned for the OpenAI family of models. With the proliferation of different models and variations, it will become more difficult to detect AI-generated content. Even if you use these tools, the result is not going to be reliable because of the wide range of statistical distributions. In some cases, it depends on the process, where AI is involved in the content creation process but not in the final step, which is done by a human. The best way to know if content was created by a machine or a human is to ask. Having policies in place in your company about the use of AI in content creation is also essential. As AI models continue to evolve, the use of detection algorithms will become less and less useful. Thank you for watching, and if you found this video helpful, please hit that subscribe button.

Summary generated by AI from the transcript.

You Ask, I Answer: Detection of AI Content?

Can’t see anything? Watch it on YouTube here.

Listen to the audio here:

Download the MP3 audio here.

Machine-Generated Transcript

What follows is an AI-generated transcript. The transcript may contain errors and is not a substitute for watching the video.

Christopher Penn 0:00

In today’s episode shall Lee asks how can people determine whether some content is created by AI or human? So this is a very popular question understandably so.

And there are some tools out there that can detect some of the language patterns have generic models.

So, for example, if you were to type into, you know, AI content detector into Google, you’ll get a whole bunch of different web services that will say like we can detect AI content and plagiarism, the all this that and other stuff.

And these tools do that to a, a greater or lesser degree of success.

The reason they can do that is because there are predictable statistical distributions in the way that large language models like the GPT family that ChatGPT uses, that are detectable that are, you can find in the generic models.

So the generic model means someone using just off the shelf ChatGPT with no customizations, no plugins, no anything, it’s just the stock base model.

And the prompt they’re putting it is so generic, that the model is essentially writing doing most of the heavy lifting.

It’s funny, these tools, these detection tools, they are iffy in their quality.

I took a blog post that Katie and I had written for the Trust Insights newsletter and I, I fed it in and there are sections that were clearly marked, this section has been generated by AI and this section is not.

And one of the tools missed everything, completely missed it.

One of the tools marked everything as AI, even the parts that we know were human written.

And then two of the tools kinda halfway winged it right some what are the tools? More or less got it? Okay.

But none of them got it right.

None of them work.

Got it perfectly right.

Nolan said, Yep, this is the AI part.

This is the non AI part.

And that’s a problem.

Right? So these tools do exist.

Their quality right now is hit or miss.

And here’s the part that’s, that’s tricky.

They are tuned for the OpenAI family of models, so GPC 3.5 GPT-4.

With the release of Facebook’s llama large language models set into open source and the proliferation of dozens if not hundreds of variations.

These tools can’t do that anymore.

Maybe schools are incapable of detecting language created by different models have different model weights, different parameters, essentially all the different settings that these other tools use that will make their texts have statistically significant distributions but different distributions than OpenAI.

And so there really isn’t a way to ironclad detect the use of AI.

The other way that these tools will fall down is depends on the process.

So if you wrote an outline as your prompt, and you had ChatGPT Write out the post, there’s a decent chance that at least some of these tools would correctly identify it.

If you did it in reverse, you said OpenAI, you write the outline because I can’t think of what to write.

And once I have my prompts as a writer, I’ll do the writing.

These tools will not detect that usage of AI even though AI was involved in the content creation process.

The final step was done by human and those statistical distributions will not exist nearly as much or as strongly as a machine generated version.

So I would say the only surefire way to know whether content was created by machine or human is to ask if it’s if it’s content that you care about.

And if it’s your content, hopefully, you would know.

But if it’s content created by your company, having policies in place as to the situations in which AI is permissible to use or not permissible to use is, is critically important because these models will continue to evolve, right? Just the open source models alone are evolving so fast and getting such specific capabilities that the plagiarism detector does not plagiarism, the AI content generation detection algorithms are going to get less and less useful.

And here’s why.

If you take an open source model, and you tune it towards a very specific task, like just writing blog posts or just writing emails, what’s going to happen is those tools will have very different language distributions.

And so something looking for the generic model is not going to see that it will just miss it.

And, again, the fact that we see see so much innovation happening on the open source side means you’re gonna have dozens, if not hundreds of models to try and keep up with.

And you’re gonna as if you were marketing, you know, AI content detection software, you’re gonna have a real hard time doing that.

So that’s the answer.

The answer is there are tools, they’re unreliable, and they will continue to get to be unreliable, they’ll actually get less reliable over time as models proliferate.

The Good question though, thanks for asking.

Talk to you next time.

If you’d like this video, go ahead and hit that subscribe button.

You might also enjoy:

Want to read more like this from Christopher Penn? Get updates here:

subscribe to my newsletter here

AI for Marketers Book
Get your copy of AI For Marketers

Analytics for Marketers Discussion Group
Join my Analytics for Marketers Slack Group!

Subscribe to My Free Weekly Newsletter

Subscribe to My Free Weekly Newsletter

Sign up now to the free Almost Timely Newsletter, released every weekend with the latest news about marketing, technology, analytics, data science, and AI.

  • This field is for validation purposes and should be left unchanged.

You have successfully subscribed to the Almost Timely Newsletter!