Why Substacks new Pangram integration is NOT a good idea.

News alert: As of this week, Substack has partnered with Pangram, a one-click AI scanner which determines whether content is human-led or AI-assisted.

While Substack has explicitly stated that “neither Substack nor Pangram uses publisher content to train generative AI models,” the partnership has already divided opinion.

I’ll hold my hands up as a hypocrite. I work in tech marketing, where engaging with AI and the discourse around it is part of my day job. I’m not in a position to preach about whether people should or shouldn’t use AI.

But in a world where human intelligence and art feels paramount, I do want to discuss why I don’t think this new Pangram integration is necessarily a good move for Substack.

To start with the intention behind the integration.

“Lets root out AI content by using an AI tool to detect it.” A noble mission right?

But what about the fact you are now using AI to analyze other people’s writing to detect AI.

And if you’re scanning content, you’ve probably had to click into it and spend time on the page. Whether or not that meaningfully affects the algorithm is up for debate, but you’re still giving attention to content that may or may not be AI.

Then what happens if the scanner flags something as AI-assisted?

There isn’t a mechanism to report it, nor remove it as AI isn’t against Substack’s rules. You could perhaps leave a comment to warn other readers, but then you’re caught in a catch 22 of further engaging with the very content you’re trying to discourage.

It’s a lose-lose situation for us. The readers and the writers.

As someone who spends multiple hours every week writing my posts for free, and spends even more hours reading to inform my content, I want to see my work rewarded over AI content just as much as anyone else.

Equally, if something I’m engaging with is using AI, I’d appreciate knowing. Especially as it’s becoming harder and harder to distinguish between human and AI-created content across every type of media.

But if the answer to achieve that is to analyze everything that exists, I don’t think that’s a worthwhile trade-off.

We will soon be entering a new era of witch hunt, with every publication and every writer open to scrutiny.

And for the sake of clarity, this isn’t about AI-produced versus AI-assisted. Every use will be demonized.

Which begs the question, how does that incentivize more people to write?

If anything, I think we’ll see more genuine writers self-censoring, filtering their writing, or choosing not to publish at all in fear of critique.

And that is one sure fire way for the AI content to win.

Whilst it for now claims that this tool is not being used to train AI models, it is a slippery slope until it is. And if this policy changes, will users be informed?

Across almost every platform or webpage, AI is quietly becoming embedded often through in-app updates that people barely notice.

How many of you have noticed WhatsApp’s AI-powered message summaries?

OR Instagram’s new AI-generated story features?

OR TikTok’s AI content labels?

These are all tangible examples of AI becoming part of content whether people asked for them or not. But its not just content platforms as you look to search platforms, shopping sites, streaming channels, finance apps, fitness trackers, call centres.

Again, I work in tech marketing and can promise you that AI can and will continue to grow.

Do I love that? Hell no.

Do I want to see us championing human intelligence instead? 100%

But is using more AI to solve the problem really the answer?

If our response to AI content is to scan, scrutinize, and question every writer, I worry we’ll create an environment on Substack that discourages the very thing we are here to support.

Many of those who I regularly engage with on here, speak about how they found Substack because they were tired of the other social platforms.

We’ve created something special, but the only way we are going to keep this community and environment is by being human and continuing with our everyday. And yes, that does mean ignoring the new AI-detection tool and anything additional that is brought in.

So I prompt you for your thoughts?

• Do you think Substacks integration of Pangram is a good or bad thing?

• And as a writer or a reader, will it change the way you engage with Substack?