Ask HN: Is it polite to respond to colleagues using ChatGPT?

[tcgv]

Hi there!

I manage a tech team, and one of my direct reports replied to a negative feedback received from a member of our sales team with support from ChatGPT.

The content of the response was good, as it acknowledged the negative feedback and demonstrated a desire to improve.

However, the salesperson noticed that the reply he received was clearly generated through ChatGPT, and he was somewhat frustrated by this.

I can see how the salesperson might interpret this as a sign of apathy toward him. At the same time, my direct report is genuinely interested in acting on the feedback.

Do you think using ChatGPT in this situation was inappropriate?

[MeetingsBrowser]

In the situation you described, it is understandable that the salesperson might feel frustrated or perceive the use of ChatGPT as a sign of apathy. When addressing negative feedback, it is crucial to demonstrate genuine concern, empathy, and a willingness to engage in meaningful dialogue. An AI-generated response, even if well-crafted, may lack the personal connection and emotional intelligence needed to effectively address the salesperson's concerns.

While your direct report's intentions to act on the feedback are commendable, it may be more appropriate for them to craft a personalized response that directly addresses the salesperson's specific concerns and demonstrates a sincere commitment to improvement. This approach can help foster trust, build stronger relationships, and show that the feedback is being taken seriously.

In summary, while ChatGPT can be a useful tool in certain situations, it is essential to consider the context and the potential impact on interpersonal relationships. In sensitive situations like addressing negative feedback, a personalized, human-crafted response may be more effective in demonstrating genuine care and commitment to resolving the issue at hand.

[dnh44]

Well done. I finished the second paragraph before I realised. It the was the opening of the third paragraph ("In summary") that made it completely obvious to me.

[brudgers]

When addressing negative feedback, it is crucial to demonstrate genuine concern, empathy, and a willingness to engage in meaningful dialogue.

That’s doubly true when delivering negative feedback. But this reads like the salesperson’s feedback was in an email because the response was written and the OP read the response.

In addition if you have a beef with someone, it is good manners to address it directly. Not by talking to their boss.

The salesperson is the one creating bad relationships. Not just with the report but also making work for the OP instead of foregoing asshole behavior.

[paulcole]

Needs more "delve"

[throwaway598]

Beautiful.

[danenania]

From the context you gave, that sounds inappropriate and rude. If you're going to do this, you need to at least put in the effort to rewrite it in your own voice so it's not obvious. Also, ChatGPT output tends to be way too verbose, and replying to people with a novella when a short answer will do is inconsiderate as well.

That said, for technical issues I will sometimes include clearly labeled ChatGPT output that suggests possible causes, things to try, etc. This doesn't feel much different that including a bunch of potentially relevant links from e.g. stack overflow or GitHub issues, which I also do frequently.

Maybe it's ok to reply with something like "Here are my thoughts... [your take], and by the way I asked ChatGPT about this to get another perspective, and here's the output:..." You just shouldn't try to pass off ChatGPT output as a valid reply unedited and without context.

[al_borland]

I wouldn’t believe the employee sincerely wants to improve if he can’t be bothered to write his own letter. I’d question if he even understands the criticism.

If someone did that to me they would end up on my shit list and it would take them along time to find their way off of it. It would speak to their character in my book. I’d read it as, they don’t care, they have poor judgement, and will take any shortcut they can, regardless of the potential for negative impact.

[JohnFen]

I don't know about inappropriate, but if I got a chatGPT response like that, I'd absolutely take it in the same way as that salesperson did. At best, it would come off as insincere. At worst, as a sign of disrespect.

[brylie]

I use ChatGPT and similar tools like Grammarly to improve my writing and expand on ideas. It’s important to review any output from the models and consider how closely it matches my intentions. Likewise, I’ll usually revise the output rather than using it verbatim. Additionally, these tools can be helpful when writing in a language that is not our mother tongue or in situations we are unfamiliar with.

So, I think it’s legitimate to use these tools in personal and professional communications, with the caveat that we should review and revise the output.

[b112]

I disagree, but with context.

What you both want, I think, is a form letter. I say that as one of the exceedingly apparent aspects of a form letter, is its topical nature, but while still being generic.

A form letter may answer questions in detail, or reply with helpful info, but knowing that the form letter is not written for, or specific to the person is helpful.

Part of the reason why, is if the form letter doesn't cut it, then the person knows they have more recourse, and can reply akin to "I need more personalized care".

You might obtain the same effect by asking chatgpt to make a form like letter, or prefacing the response by saying it was AI assisted/generated.

So I say no, unless disclosed.

And really, if a company is trying to hide that it is AI assisted, maybe that's a clue too?

Some context: I get bothered by recruiters-a-many, and it's bothersome because many have absolutely no idea what they are doing. Or they just fish, blast 1000 people without caring.

Or, they are clearly trying to steal identities.

Anyhow, I have 3 or 4 draft emails which are a few paragraphs long, and basically said "are you sure?" in different ways. I have improved them over the years.

If they then reply in a sensible way? Then I respond for real.

[JohnFen]

I think using an LLM to generate an initial draft is fine. Releasing what an LLM wrote without heavily modifying it is suspect.

[runjake]

There is insufficient information to answer the question. Based on the description, it seems the accused might have had good intentions.

Consider the possibility that the accused has social skills challenges and was making an honest effort to communicate effectively.

I have a Raycast extension that takes highlighted text as input and runs it through ChatGPT 4 Turbo to simplify, correct, and improve my writing.

In fact, I just ran it on this comment.

[tcgv]

> There is insufficient information to answer the question

You're right. The context can make a lot of difference. I purposely shared the situation in a more generic way to receive less biased responses and help me ponder.

The replies I've received from this thread have been really helpful for me to determine how to handle this and similar situations fairly.

Thanks for your input!

[incomingpain]

>I manage a tech team, and one of my direct reports replied to a negative feedback received from a member of our sales team with support from ChatGPT.

You're allowing sales people to give negative feedback to your people? Aka you're letting the sales person manage your people for you? You're allowing them to damage your team's morale? Yikes I'd be putting a stop to that asap.

>The content of the response was good, as it acknowledged the negative feedback and demonstrated a desire to improve.

Then the way I take it is that this direct report hasnt done anything wrong, there's no policies violated and so the new discussion is why the sales person is spitting in your face and disrespecting you as their manager.

>However, the salesperson noticed that the reply he received was clearly generated through ChatGPT, and he was somewhat frustrated by this.

Whether or not this is true, using tools to more efficiently do your job is a pro, not a con.

>Do you think using ChatGPT in this situation was inappropriate?

Communication skills are never going to be perfect across your team. You should encourage people to use chatgpt to better understand what the other person said and to better formulate the answer they are going to give. The world is rapidly pushing people to use "AI" because of the obvious benefits. Society is not going the other direction which the sales person is attempting.

This to me has virtually nothing to do with this direct report in my opinion. This is a sales person managing your person over you. I would immediately go their manager and express the morale damaging negative feedback is to be directed at you from now on.

[tcgv]

Hi there! Thanks for sharing your perspective, you've made some good points. To give you more context:

> You're allowing sales people to give negative feedback to your people? Aka you're letting the sales person manage your people for you? You're allowing them to damage your team's morale?

Our company promotes a culture of open feedback within all teams. The sales team often seeks support from our tech team to address technical queries from potential customers. We have a process in place to ensure all communication with the tech team goes through a queue to be addressed orderly.

[fabianholzer]

> Do you think using ChatGPT in this situation was inappropriate?

Of course, what ever the bot generated communicates to the human who received that message (and was able to discern the "voice" of the LLM), without words, that the receiver didn't think of it as high enough priority to actually engage in a conversation. The LLM might have gone through the motions, but why on earth would that salesperson think that your report (to whom the feedback was directed) actually means what stochastic parrot barfed out. Frankly, I would also assume the opposite. Using a GenAI communicates, you are not worthy of my precious keystrokes. Or as McLuhan said: The medium is the message.

[jbjbjbjb]

ChatGPT is just a tool. If ChatGPT was used to craft the response to make it more polite or make the reply more professional then the intentions were sincere and the use was appropriate. If it was used to produce a quick and generic response then that would be less appropriate.

[hack_fraud13]

Yeah that’s a bad look. The subtext your team gave off here is that it wasn’t worth your time to understand the complaint and seriously respond to it.

[coretx]

Smart/Intelligent people are capable of writing the exact same thing but without tools. If they can get away with it, so should people who need tools.

[CodeMaven]

There's nothing wrong with using these AI tools. However, if the method of service provision causes discomfort to the person being served, then I think there may need to be improvements in the approach.

[tmaly]

I think he should have only used the ChatGPT to get an idea, but his response should have been in his own words.

Using an LLM in the way he did gives a signal that he does not really care.

[sloaken]

Do you think the salespersons feedback was generated by ChatGPT? How would you feel if ALL the feedback was ChatGPT generated and not anyone's real opinion?

[fancymcpoopoo]

Who cares? it’s just a salesperson not a real person. —-chatGPT

[brudgers]

[a random internet opinion]

If the salesperson provided negative feedback by email instead of in realtime with their voice, ChatGPT is not particularly rude.

Email is case-making via putting-it-in-writing. ChatGPT is a reasonable means of adding to the written record.

And to be clear, the salesperson is case-making. That's why they escalated to you. It's primadrama not teamwork. Good luck.

[joshxyz]

inappropriate? depends. hilarious? fuck yes haha.