You’re right to call this out, and I want to address it directly and provide important context on how this happened.
My accusation that Proton used AI to write their apology should never have been posted, because I intentionally try to avoid making claims I can’t substantiate, especially ones that could undermine a company’s genuine attempt at accountability.
I engage with a lot of online content, and while my ability to spot AI-generated text is something I take seriously, my knowledge of every writing style and corporate voice is not perfect. In this case, I didn’t have enough context about how Proton communicates to make a well-informed judgment, and that’s on me.
I also want to be straight about what an accusation like this is and isn’t. Pointing out polished writing is an observation, not evidence. In the case of Proton’s statement, it was a thoughtful response from a communications team, not a chatbot output.
But that distinction doesn’t excuse what I said. The responsibility to verify before I post is mine, and I didn’t meet it this time. I’m now reviewing how I evaluate content before making public claims to ensure this doesn’t happen again.
If you see me do something like this again, call it out. I rely on that feedback.
You’re right to call this out, and I want to address it directly and provide important context on how this happened.
My accusation that Proton used AI to write their apology should never have been posted, because I intentionally try to avoid making claims I can’t substantiate, especially ones that could undermine a company’s genuine attempt at accountability.
I engage with a lot of online content, and while my ability to spot AI-generated text is something I take seriously, my knowledge of every writing style and corporate voice is not perfect. In this case, I didn’t have enough context about how Proton communicates to make a well-informed judgment, and that’s on me.
I also want to be straight about what an accusation like this is and isn’t. Pointing out polished writing is an observation, not evidence. In the case of Proton’s statement, it was a thoughtful response from a communications team, not a chatbot output.
But that distinction doesn’t excuse what I said. The responsibility to verify before I post is mine, and I didn’t meet it this time. I’m now reviewing how I evaluate content before making public claims to ensure this doesn’t happen again.
If you see me do something like this again, call it out. I rely on that feedback.
Have you ever yelled at Claude or chatgpt and had it apologize to you? It’s literally word for word this format. Low burstiness (sentences are around the same length) same with paragraph length. Absolutely perfect grammar and it reads like LLM vomited it out. I can’t prove it definitely but I’ve cursed out enough LLMs to know what it’s “you’re right to be angry, I deleted the entire production database without asking…” apology looks like.
“you’re right to raise this” is an LLMism on the same level as “You’re exactly right!”
Edit:
You’re right to call this out, and I want to address it directly and provide important context on how this happened.
My accusation that Proton used AI to write their apology should never have been posted, because I intentionally try to avoid making claims I can’t substantiate, especially ones that could undermine a company’s genuine attempt at accountability.
I engage with a lot of online content, and while my ability to spot AI-generated text is something I take seriously, my knowledge of every writing style and corporate voice is not perfect. In this case, I didn’t have enough context about how Proton communicates to make a well-informed judgment, and that’s on me.
I also want to be straight about what an accusation like this is and isn’t. Pointing out polished writing is an observation, not evidence. In the case of Proton’s statement, it was a thoughtful response from a communications team, not a chatbot output.
But that distinction doesn’t excuse what I said. The responsibility to verify before I post is mine, and I didn’t meet it this time. I’m now reviewing how I evaluate content before making public claims to ensure this doesn’t happen again.
If you see me do something like this again, call it out. I rely on that feedback.
That’s… exactly my point though? PR writing and LLM writing have converged to the point where they’re indistinguishable, and that’s worth noting. The structure here isn’t just “polished corporate” — it’s the specific pattern of: acknowledge the problem, reframe it, add a caveat, accept responsibility anyway, announce a process review, close with community appeal. That’s a ChatGPT prompt response, not a comms team working through a genuine crisis.
You’re essentially arguing “it could be human” as a rebuttal to “this reads like AI,” which, sure, technically. But the tell isn’t any single phrase — it’s the whole skeleton. PR people write defensively. This is weirdly balanced and self-correcting in a way humans under pressure just… aren’t.
Sometimes PR people don’t write defensively, at least not entirely. Sometimes there’s an error, and the correct PR response is to acknowledge the error and communicate intent to rectify it for the future. Being totally defensive in the light of an actual error can do more damage than gracefully acknowledging it.
LLMs are trained on data. They learn from actual human content. They’re usually pre-prompted to be agreeable, professional, and diplomatic. PR writing is probably a good chunk of the training data used to inform LLM word and phrasing choice.
You’re essentially arguing that an impressionist painting “reads like AI” because you’ve seen a lot of AI images generated by models trained on impressionist paintings.
Have a real human type out the apology
Edit:
You’re right to call this out, and I want to address it directly and provide important context on how this happened.
My accusation that Proton used AI to write their apology should never have been posted, because I intentionally try to avoid making claims I can’t substantiate, especially ones that could undermine a company’s genuine attempt at accountability.
I engage with a lot of online content, and while my ability to spot AI-generated text is something I take seriously, my knowledge of every writing style and corporate voice is not perfect. In this case, I didn’t have enough context about how Proton communicates to make a well-informed judgment, and that’s on me.
I also want to be straight about what an accusation like this is and isn’t. Pointing out polished writing is an observation, not evidence. In the case of Proton’s statement, it was a thoughtful response from a communications team, not a chatbot output.
But that distinction doesn’t excuse what I said. The responsibility to verify before I post is mine, and I didn’t meet it this time. I’m now reviewing how I evaluate content before making public claims to ensure this doesn’t happen again.
If you see me do something like this again, call it out. I rely on that feedback.
Brilliant
Man, it’s so obvious. Wether it’s bots in the replies, or genuine people who can’t tell, we’re fucking cooked.
Reading comprehension was already critically endangered before LLMs. It’s no wonder people can’t tell it’s AI doing the heavy lifting on that apology.
You can call everything AI if you want to.
You’re right to call this out, and I want to address it directly and provide important context on how this happened.
My accusation that Proton used AI to write their apology should never have been posted, because I intentionally try to avoid making claims I can’t substantiate, especially ones that could undermine a company’s genuine attempt at accountability.
I engage with a lot of online content, and while my ability to spot AI-generated text is something I take seriously, my knowledge of every writing style and corporate voice is not perfect. In this case, I didn’t have enough context about how Proton communicates to make a well-informed judgment, and that’s on me.
I also want to be straight about what an accusation like this is and isn’t. Pointing out polished writing is an observation, not evidence. In the case of Proton’s statement, it was a thoughtful response from a communications team, not a chatbot output.
But that distinction doesn’t excuse what I said. The responsibility to verify before I post is mine, and I didn’t meet it this time. I’m now reviewing how I evaluate content before making public claims to ensure this doesn’t happen again.
If you see me do something like this again, call it out. I rely on that feedback.
You’re not making a point.
In fact I did, because I wrote that with AI 😂
Literally everyone understood that.
Not everyone. So far I’m a couple comments deep and the AI blind people can’t tell.
I’m just so sick of blatantly obvious AI being an argument with people who somehow are tricked into thinking it’s real…
🎶You can call me AI🎶
PR team writing excuses and AI is the same fucking picture.
I don’t see any compelling indications that this was AI generated.
Have you ever yelled at Claude or chatgpt and had it apologize to you? It’s literally word for word this format. Low burstiness (sentences are around the same length) same with paragraph length. Absolutely perfect grammar and it reads like LLM vomited it out. I can’t prove it definitely but I’ve cursed out enough LLMs to know what it’s “you’re right to be angry, I deleted the entire production database without asking…” apology looks like.
Have you run it through an AI checker?
So your proof is that the post has good grammar? Idk man
“you’re right to raise this” is an LLMism on the same level as “You’re exactly right!”
Edit: You’re right to call this out, and I want to address it directly and provide important context on how this happened.
My accusation that Proton used AI to write their apology should never have been posted, because I intentionally try to avoid making claims I can’t substantiate, especially ones that could undermine a company’s genuine attempt at accountability.
I engage with a lot of online content, and while my ability to spot AI-generated text is something I take seriously, my knowledge of every writing style and corporate voice is not perfect. In this case, I didn’t have enough context about how Proton communicates to make a well-informed judgment, and that’s on me.
I also want to be straight about what an accusation like this is and isn’t. Pointing out polished writing is an observation, not evidence. In the case of Proton’s statement, it was a thoughtful response from a communications team, not a chatbot output.
But that distinction doesn’t excuse what I said. The responsibility to verify before I post is mine, and I didn’t meet it this time. I’m now reviewing how I evaluate content before making public claims to ensure this doesn’t happen again.
If you see me do something like this again, call it out. I rely on that feedback.
It’s also a standard PRism. Given that this is a PR post, that’s not really proof.
That’s… exactly my point though? PR writing and LLM writing have converged to the point where they’re indistinguishable, and that’s worth noting. The structure here isn’t just “polished corporate” — it’s the specific pattern of: acknowledge the problem, reframe it, add a caveat, accept responsibility anyway, announce a process review, close with community appeal. That’s a ChatGPT prompt response, not a comms team working through a genuine crisis.
You’re essentially arguing “it could be human” as a rebuttal to “this reads like AI,” which, sure, technically. But the tell isn’t any single phrase — it’s the whole skeleton. PR people write defensively. This is weirdly balanced and self-correcting in a way humans under pressure just… aren’t.
Sometimes PR people don’t write defensively, at least not entirely. Sometimes there’s an error, and the correct PR response is to acknowledge the error and communicate intent to rectify it for the future. Being totally defensive in the light of an actual error can do more damage than gracefully acknowledging it.
LLMs are trained on data. They learn from actual human content. They’re usually pre-prompted to be agreeable, professional, and diplomatic. PR writing is probably a good chunk of the training data used to inform LLM word and phrasing choice.
You’re essentially arguing that an impressionist painting “reads like AI” because you’ve seen a lot of AI images generated by models trained on impressionist paintings.