ChatGPT and Nonprofit Succession – AI’s Not Quite There
Photo by Rock'n Roll Monkey on Unsplash
Curious to play with AI, I typed “Write a blog post about succession planning in nonprofits” into ChatGPT. With some irritation, I watched it churn up a 542-word post in less than 30 seconds. ChatGPT wins, hands down, at speed. But do we agree with it? Laurie and I gave it a critique.
Here’s an example of what we like:
“Succession planning in nonprofits goes beyond finding a replacement for a departing leader; it's a strategic process that safeguards the organization's mission, values, and continuity.” Yes, AND it also makes the team stronger before and after any transition.
“Leadership changes can be disruptive, but with a succession plan in place, the transition can be smoother, minimizing confusion and maintaining operational efficiency.” AND this is especially true if those operating efficiencies are generated by a coordinated team.
But these examples are dubious:
“Successful nonprofits are built upon the dedication and expertise of their leaders.” Wait, what about the rest of the staff and volunteers? Are you saying they can just stay home?
“A thoughtful succession plan ensures that this institutional knowledge is transferred to new leaders, preventing loss of expertise.” Are “leaders” the only one with knowledge? You’re doing it wrong if so.
ChatGPT is entirely missing any reference to equity, inclusion, or even simple diversity. The entire case for succession planning is presented from the perspective of the leader, the organization, and stakeholders including “donors, volunteers, and partners.” What about the people the nonprofit serves? What about the staff and their aspirations?
I asked ChatGPT how it works:
“In essence, ChatGPT learns the patterns and structures of language from its pre-training phase and then fine-tunes this knowledge for generating human-like responses in conversations during the fine-tuning phase. The model's effectiveness largely depends on the quality and diversity of the training data, as well as the specific fine-tuning process.”
So, if the “training data” that ChatGPT uses comes from a society with systemic racism, it follows that the output will parrot that thinking – at least initially. Our takeaway is that ChatGPT may provide a better history lesson than acting as a forward-thinking tool.