my angry boss uses AI to write kinder emails … and it feels weird

A reader writes:

My manager, Athena, has pretty poor soft skills and often comes across as aggressive, interrogating, micromanaging, and dismissive. This happens both in person and over email and instant message. In writing, her spelling and grammar are also inconsistent and her phrasing is often curt or abrupt.

Lately, my coworkers and I have noticed a huge improvement in some of her emails and chat messages. Emails and messages that previously would have been curt and aggressive are now warmer and softer, with perfect spelling and grammar. It’s theoretically the exact change her direct reports have all been desperately wishing for, except that her in-person communication has not changed. Her tone is still angry, her approach is still aggressive, and feedback discussions still feel like an interrogation.

We’re pretty sure she’s using AI to rephrase her written communication, because she’s talked openly about using Claude to help with writing emails and project reports. It’s also not consistent — most of her text communication is the same as before, with maybe 30-40% being the new and improved version.

Even though it’s technically better to be getting at least some non-aggressive emails from our boss, it feels weird and disingenuous. On one hand, it’s good that she recognizes a need for improvement, and it’s a relief that some of her emails sound like they come from the pleasant, patient manager we wish we had. On the other hand, the contrast is so obvious that it feels like putting a false veneer over the deeper problem of what she’s really like. As one coworker said, “I get email from Athena or I get email from Claude.”

I know there’s not much we can do either way, but are my coworkers and I justified in feeling creeped out and vaguely insulted by the clearly AI-generated emails and texts we’ve been getting? Would love to hear your take on the ethics and optics of this sort of AI use.

I don’t know, this feels like at least a partial win to me. You used to get emails from negative, dismissive Athena but now you (at least sometimes) get emails from kinder, more socially appropriate Claude.

The problem is that negative, dismissive Athena is still manages you the rest of the time.

I’m actually really interested to know if people like Athena who use AI this way will over time start to learn how to adopt a warmer tone themselves. I have long been convinced that people who default to Athena-like communication genuinely don’t know how to envision what different language or a different tone would sound like. They think being warmer means making lots of disingenuous chit-chat (when that’s not at all what it needs to mean) or that they’d have to sugarcoat everything to the point of it becoming meaningless (also not what it means). And so over time, there might be significant learning advantages to her seeing what Claude does to her communications. Or not, who knows. But I’d be really interested to watch how that plays out.

You see the opposite of this, too: people who are very passive and indirect in their communications can’t picture a healthy, assertive version of their communications; instead, what they picture in their heads feels confrontational (often because the models they had for conflict growing up were very bad ones). It’s one reason why I try so much to give sample language here, because I think there’s real value in saying, no, it could just sound like this. I have many, many ethical issues with generative AI, but since it’s here, I am very interested to see if using it in this way over time can help people envision better language on their own at some point.

Anyway … is it disingenuous of Athena to be using AI to revise her emails? I don’t know that it is! I’d call it that if she were rolling her eyes while sending them and thinking, “These delicate flowers require such careful handling.” But if she’s running her language through AI and thinking, “Okay, that sounds good, I’ll send that,” I don’t think it is disingenuous.

But I can also understand why it feels that way to you, especially since the rest of her communications have remained the same old dismissive Athena.

Ultimately, your problem is that you have a manager who’s angry, interrogating, micromanaging, and dismissive, not that she sometimes uses a tool to improve her tone. Is there anything you can do about that, like having a discreet word with someone above her? If not, I’d try to embrace the times where she’s at least letting Claude help her — or, if that’s not possible, at least see the humor in the extreme contrast between her and her robot assistant.

The post my angry boss uses AI to write kinder emails … and it feels weird appeared first on Ask a Manager.

View the full article