Annoyance, Grief, and Existential Dread (Oh My!)

This week, I followed an AI policy from draft to approval, made some progress on a tool to help me track what happens next, and worked my way from mild annoyance with AI to complete existential dread. So, a pretty normal week.

Policy Dive, Continued

The first thing I did this week was confirm that SDCOE’s AI Policy (BP 0441) was, in fact, approved. In last week’s AI in Ed policy dive, I noted that SDCOE had created a draft policy, but I could not tell from the published documents whether it had actually been approved at the 9/9/26 meeting.

So I watched the 9/9/26 meeting video.

The brief discussion confirmed the policy’s approval, but it also referenced the first reading on 8/12/26, when board members had a much more substantive conversation and suggested changes that helped shape the final version.

So I watched the 8/12/26 video, too.

During the first reading, board member Dr. Erin Evans suggested several small but important clarifications to the policy. The board seemed generally in agreement with those changes, and the discussion that followed raised several larger questions:

• Who approves AI tools?
• Who (read: humans) verifies the research behind these tools, especially when AI could also be used to conduct it?
• How should a policy address privacy when many AI tools are controlled by large, profit-driven technology companies and may share data with them?
• How do you write policy for technology that is changing much faster than policy ever could?

What emerged from the discussion was that BP 0441 is aspirational. It is a statement of the values and expectations SDCOE has now, with the understanding that the policy will need to evolve as the technology, research, and evidence do.

Dr. Evans also referenced a few items worth reading:

• The Students First Act, an AI policy for education created by students themselves. NPR did a nice article about it.
• Brookings Institution: A New Direction for Students in an AI World: Prosper, Prepare, Protect.

There were plenty of other threads I wanted to pull at this week, but this is not my day job, so I only got so far.

Docket

One way I’m trying to make sense of all of this is by building a better way to track it. I was able to make some preliminary progress on an AI-assisted screening tool I’m currently calling Docket.

The idea is to build a weekly dashboard to help me track AI policy across SDCOE and the six San Diego school districts that enroll about half of the county’s students.

For phase one, I want to track AI-related items from the moment they show up on a school board agenda through whatever actually happens next.

Below the main queue are a few other views I’m experimenting with: the week’s meetings, the policy threads that might eventually turn into essays, and a coverage grid showing where I’m finding activity and where I’m not. This is very much a first iteration. I am trying to use AI to make the screening and tracking faster, allowing me to decide what is actually relevant and accurate. Here is what it looks like so far.

I also created accounts on X and Threads to keep up with news and developments.

Annoyance, Grief, and Existential Dread (Oh My!)

While I was building tools to keep track of what was happening locally, the broader conversation around AI seemed to be taking a darker turn. It felt like public discourse crossed a tipping point into anti-AI sentiment. Or maybe it was always there and I just started paying attention?

The week started off with the leaders of Anthropic and OpenAI calling for a coordinated slowdown in AI development. They shared how they have caught people using AI for evil, and AI scheming.

I started the week experiencing annoyance at the imperfections of AI in editing my writing. I spent about an hour talking through edits on my last post. ChatGPT discussed the changes thoroughly, but it failed to make many of the edits it claimed to be making. I had to go back and repeat an hour’s worth of work when I was already tired of staring at the screen. I noticed this happened a few times in my coding work as well. Claude would make new changes, then I would discover that older changes had been undone or incomplete work had been pushed prematurely. I had to diagnose the problem and roll everything back.

As the week went on, my annoyance shifted to existential dread.

Developers I trust, including people who understand AI’s potential better than most, were sounding the alarms. The conversation shifted from “How can we harness the benefits of AI?” to “We need to control AI to ensure the survival of civilization.”

And I began to wonder if all of this had thrown me into the grieving process. I was feeling my way through a cloud of cognitive dissonance, experiencing both what AI allowed me to do and my growing dependence on it, while also accepting that, for all its superpowers, AI is not infallible and its downside risks could be catastrophic.

But the grieving process does not quite capture what is going on. So I did some “research.” I asked ChatGPT (facepalm), “Are there models of human psychology for adapting to new threats, similar to the grieving process?” It came back with six potentially relevant theories and provided a useful synthesis:

Disruption → threat appraisal → emotional response → attempts to regain control → experimentation/coping → revised expectations → adaptation.

Last week, I seemed to move from threat appraisal to emotional response. This week, I expect to focus on regaining control and coping through experimentation. I have already found myself looking into ways I can protect my data and thinking about how I’d feed my family if the power went out.

As with most psychological models, these aren’t strict, sequential stages. We can expect to loop back, repeat, and skip stages in our usual messy human ways. No matter what happens this week, I’ll be sure to write about it here.

Until next time, stay intelligent, humans.