Let’s talk about AI.
If you’ve scrolled all the way to the very end of my regular newsletters anytime in the past (I’m not sure how long 😂), then you already know that I have a note in the footer that this is 100% hand-crafted content, without even a drop of AI.
I don’t use AI for any part of my work – not for brainstorming, not for creating a first draft, not for checking my final draft for “tone”, not for any part of my process. I’ve changed to using different search engines since Google started pushing their AI summaries. I disabled it in my Firefox and gmail settings, and anywhere else that I can. I know I can’t avoid it completely – I use some social media platforms that use AI to power their algorithms – but I avoid it as much as I can.
We all have different reasons for the choices we make around whether or not to use LLMs. As we’ve learned more about the effects of AI – from providing deadly misinformation, to high profile AI psychosis cases, to the environmental impact of data centers, the plagiarism, and on and on – I’ve come to the conclusion that using AI is not in alignment with my values. I’m happy to talk more about that if you’re interested, but most of those reasons are beside the point of this newsletter!
So here are the main issues I have with AI in relation to leadership and coaching:
Does it really save you time?
That’s one of the big arguments in favor of using it, right? It’s supposed to do so much of your work for you!
When I talk with library leaders about embracing their curiosity and using a coaching approach with their team, one of the biggest barriers is time. You have so much on your plate, that you feel pressured to just solve the problem, instead of taking the time to develop your team member’s ability to solve the problem, so that they can take full ownership of that work.
So if AI were actually able to do the tedious parts of your job reliably and accurately, so that you have more time for building relationships and developing your team members, then that would be a point in its favor.
Unfortunately, I’ve mostly heard stories of how it’s shifting that burden to fact checking, fixing errors, and avoiding hallucinations, and actually taking up more time than it’s saving.
How is it affecting your relationships?
The one place I’ve heard people say it’s saving them time is in dealing with email. As someone who can sometimes spend half an hour writing and rewriting a 3 sentence email, that sounds great! But, that can erode your human relationships.
People are growing increasingly vigilant about detecting AI writing – including maligning my beloved em dash as a “tell” 😭
Research has found that managers who use AI to write work emails are perceived as less trustworthy. You don’t have to look far online to find anecdotal evidence that even mild use, like having Claude adjust the tone of your emails, can creep out some employees. Even Ask A Manager has a post responding to this issue.
So regardless of your opinion of using AI in general, I’ll encourage you to think critically about what message it can send to your team members if you choose to automate your relationships with them.
How is it affecting your curiosity?
The reason I decided to write about AI today is because I read some recent research findings:
Researchers … found that access to AI advice collapsed people’s willingness to say “I don’t know” from 44% to 3%. Accuracy dropped from 27% to 9%. Confidence, meanwhile, rose from 30% to 76%.
(
Source)
Being willing to say “I don’t know” is fundamental to curiosity and learning.
The researchers were interested in the cognitive processes involved and people’s willingness to suspend judgment. The experiments were engineered to ask questions that have objectively correct and incorrect answers. The AI agent was designed to give wrong answers most of the time, so that reliance on this AI agent couldn’t be justified by it having any track record of providing accurate answers. The results were so stark that the researchers suspect that people who would have answered correctly on their own chose to give the AI agent’s answer instead.
People usually undervalue advice from other people, but participants overvalued advice from the AI agent in this study.
It would be one thing if this only happens when being asked to give a simple factual response, like the set-up in this research. But the researchers see no reason to expect this cognitive effect to only apply to this specific context.
If you do choose to use AI to brainstorm or to ask for other perspectives on a situation, make sure to only use that as a starting point for your own curiosity. Don’t give the AI response any more weight than you would give to advice from some random stranger.
And make a point to ask questions and actively listen to your team members, no matter what advice your AI has already given you. Your team members know more about your specific context than any AI possibly could.
It’s not challenging you to be better.
Research on several different AI agents has found that it tends to be sycophantic, telling you that your ideas are great, no matter how bad, harmful, or even illegal they may be. They’re programmed to make you like them, so that you’ll keep using them. That means that they’re not going to challenge you to be better, and can actually make you worse at leading a team.
Just like Google was never going to replace librarians, Claude (or any other AI) is never going to replace human coaches. Coaching is fundamentally about challenging you to question your assumptions and consider a wider range of perspectives. I don’t generally tell anyone “that’s a bad idea”, but I do ask what it was that led you to that idea, and then dig in more from there.
Whether you’re ready to work with a professional coach toward a larger goal or just need a sounding board to talk through how to approach one difficult conversation, please talk through it with a real, live, human being, not any AI agent.
My takeaways
I won’t judge you here if you choose to use AI for some things, but I will highly encourage you to think critically about when, where, and how you use it.
Effective leadership is about working with people – building relationships, building trust, learning, growing, and helping your team members develop. Please don’t let AI get in the way of your ability to accomplish these things.
What are your takeaways?
Drop me a reply to let me know what stood out for you and/or what points you want to push back on.
And keep scrolling for your invitation to join a real live human discussion about leading with curiosity on Friday, Aug 7, no AI bots allowed!
Some good news
In case you haven’t seen this article yet:
Librarians are hosting viral ‘Avoiding AI’ workshops for people who are fed up with Big Tech
And they’re drawing huge crowds to their libraries for these workshops!
Leave a Reply