Summary

The Medium platform is currently facing an influx of unacknowledged AI-generated content, leading to a call for community vigilance and the use of AI detection tools to combat the issue.

Abstract

The article discusses the prevalence of AI-generated content on Medium, which is causing frustration among human authors who invest significant time and effort into their writing. Despite Medium's stance on being a platform for human storytelling, the current guidelines do not effectively prohibit or penalize the posting of AI-generated content. The community is urged to take action, as relying on Medium's policies alone is insufficient. The author suggests that human detection methods are subjective and time-consuming, and instead proposes the use of AI detection tools as a more efficient solution. These tools are described as being in an arms race with AI content generators, constantly evolving to improve their accuracy in identifying AI-generated text. The author plans to review top-performing AI detection tools in a follow-up article and invites community input on tools they currently use.

Opinions

The author believes that Medium's current policies are inadequate for addressing the issue of AI-generated content, as they do not impose significant penalties for posting such content.
There is a sense of injustice among genuine authors who feel that their hard work is being undermined by AI-wielding "grifters" who flood the platform with machine-generated content.
The author points out the impracticality of relying on subjective human judgment to detect AI content, emphasizing the time-consuming nature of such efforts and the potential for misinterpretation.
The author is optimistic about the potential of AI detection tools to level the playing field, noting their objectivity and efficiency compared to human detection methods.
The author values community engagement and invites readers to contribute their experiences with AI detection tools, indicating a collaborative approach to tackling the problem.

The AI vs. Human Arms Race

Medium Is Not Going To Save Us in the Fight Against AI

Undisclosed AI scammers presently have the upper hand. We can fight back, but we have to be smart about it ourselves

**Is that a soulless gaze or just AI content?** | Image James Bellerjeau

If you suspect a lot of content you’re seeing is unacknowledged AI-generated text, you’re right.

So many of us pour our hearts and souls out onto the page, only to see shameless grifters flooding the field with machine garbage.

It drives hard-working Medium authors crazy, and with good reason. With a fraction of the time and effort we invest, AI-wielding thieves are stealing away attention.

Some of you are thinking, “Wouldn’t it be nice if Medium did some of the work of spotting and blocking these raging opportunists for us?” Yes, of course. But Medium has abdicated responsibility.

Although management says Medium is “a home for human writing, full stop,” we’ve got more moochers living under our roof than the Delta House fraternity.

Medium’s statement is toothless because there is scant penalty for posting AI-written or AI-assisted stories.
Medium’s ironically named “Quality Guidelines” do not prohibit unlabeled AI-generated stories, merely restrict their broad distribution.
To get around even this feeble mechanism, all an AI hack needs to do to stay in Medium’s good graces is say upfront they used AI.

So for now it’s up to us, the Medium community. For those of us who want to fight back, we can fight back.

The only question is how best to do it. I’ve got bad news and good news.

We ain’t going to fix this with our gray matter, sorry

Several fine authors have written guidance on how to tell whether an author is posting AI-generated content. Thanks, R C Hammond!

The tips are good but suffer from two fatal flaws:

They’re inherently subjective
Even when they work, they’re massively time-consuming

If you write a lot here, you’ve probably been wrongly accused of using AI yourself. Or maybe rightly. Either way, talk about annoying!

The point is that with a diverse enough group, which Medium certainly contains, reasonable people are going to interpret the same facts differently. The po-TAY-to | po-TAH-to schism is real.

Yes, if you are a super-sleuth, and devote 30 minutes or more to scouring a writer’s profile, checking their posts and comments, you might gain confidence that they’re a con.

Who has the time? And why should we devote so much time to rooting them out? Especially when their remedy is to take five minutes to set up another fake profile and continue on their merry, malicious way?

This is to say nothing of people who use AI for all sorts of legitimate reasons, blurring the lines further. See Wesley van Peer’s excellent rant laying out another side of the story.

**Grandma Smillew, patiently meditating on AI** | Image James Bellerjeau

It takes a bot to catch a bot

Humans are losing the current battle, but we haven’t lost the war. We just need to employ more powerful tools against the powerful tools arrayed against us.

AI detection tools to the rescue.

We’re in the raging arms race phase of this fight. Ever more sophisticated AI-generated text versus AI detection tools.

Do you know the big difference between AI detection tools and our fallible brains? The tools are not subjective. And they work instantly.

The AI detection tools are trained the same way that large language models are, on huge databases of content. Then, and this is critical, their performance is checked and fine-tuned against known AI-generated content.

In other words, there’s no guesswork when training these tools. The fakes are salted in and there’s no question what is AI-generated. The only question is how to make the tool catch fakes more and more reliably.

In Part 2 of The AI vs. Human Arms Race (coming soon), I review the top-performing AI detection tools.

Do you use any AI detection tools now?

Drop a note in the comments and I’ll make sure to include the tool in the review.

Be well.

Whether you’re a Luddite or a closet cyborg, I love you all. Subscribe to read my stories, and you’ll gain a human fan who is bot-like in his determination to engage with you.