Welcome to our February 2026 Newsletter!
I’m hoping you find these newsletters useful. If you have feedback, you can email me directly at kkrugler@aiforcommunity.org. And thanks for reading!
Updates on AI for Community
There are some easy things you can do that would help us hit our goal next year of teaching over 1000 nonprofit staff & volunteers about AI. And we’re continuing to expand our courses…
- We are now hosting regular online meetups. The last one was February 12th, where we talked about using AI for grants (video here) – thanks to everyone who participated, that was fun. The next one is March 12th, 10am PST, where we’ll be talking about AI policies (register here).
- We’ve started planning for a more advanced class, which covers customizing your AI, deep research, agents, and AI policies. Details coming soon in our next newsletter!
- If you know people who would benefit from the training, point them at our registration page. The next one is February 23rd. We also have a March 18th class available.
- If you know anyone who works at a community foundation or large nonprofit, please introduce them to us. We need to partner with organizations to efficiently set up training. An easy way is to have them schedule a Zoom call with me.
Hot Off the Press
Quite a lot has happened over the past month…
- OpenAI released ChatGPT-5.3, which is much better at “deep thinking”, and now feels similar to Gemini 3 Pro in terms of capabilities (other than Gemini handling much larger documents).
- There’s a cheaper version of the paid ChatGPT, called “Go”. It’s $8/month. You get the ChatGPT-5.2 Instant model, which is still pretty good.
- The free and “Go” version of ChatGPT will likely have ads soon. The Claude commercials during the Super Bowl show why this might not be great, and were pretty funny.
- OpenAI updated Sora (its video generation system). It can now do a solid job of creating a video from an image (check out the results of “kkrugler enthusiastically teaching AI to a room of nonprofit staff”).
- Google added “Personal Intelligence” to Gemini. This creates a pretty seamless integration between Gemini and your Google Workspace-based email, calendar, etc. So you can say things like “Summarize the board meeting minutes from Drive”.
- Anthropic rolled out Opus 4.6, which is now (to me) clearly the best model for general writing and strategy/planning.
- Anthropic also announced that free tier users can now create files, use connectors, and access skills. These are more advanced features that previously required a paid plan.
Tip of the Month
Something I talk about in training is how useful it can be to have free-form text fields in survey forms. It’s now possible to have AI do a reasonable job of analyzing this text aDuring class I’ve talked briefly about how to customize AIs to improve results. Below is a brief list of customizations I’ve done with Claude (some of these require a paid subscription):
- In Settings > General, I set my Personal Preferences to read: Prioritize accuracy: verify facts, cite sources when available, and challenge questionable claims or assumptions—mine or yours. End responses with a confidence score and brief justification.
- In Settings > Capabilities, I turn on everything. That includes Code execution and file creation, Search and reference chats, and Generate memory from chat history.
- In Settings > Connectors, I enable Google Drive, Gmail, and Google Calendar. These are all for my work (AI for Community) account, NOT my personal account.
- In Settings > Organizational Settings > Capabilities, I turn on everything. This includes Web search, Interactive content, Enable artifact connectors, etc.
- In Settings > Organizational Settings > Claude in Chrome settings, I enable it for all team members.
And then I create artifacts that describe my organization, for example a “Quick Facts” overview, and a more in-depth “Grant Evaluation Profile” that contains details relevant to assessing funding opportunities.
Final Thoughts
A year ago, when I first started teaching these classes, the hot topic was “prompt engineering”. As in how to write prompts that worked well when iterating with AI. And that’s still important, but as AIs have gotten better, the focus is shifting to what I’ll call “context engineering”. AI has gotten pretty good at figuring out what you mean in a prompt, and how to work with you to clarify your goals. But without good context, the results are often going to disappoint. So I find my focus shifting to “what relevant documents do I have, and what do I need to create”, as a key first step in any AI-based activity. And I’m finding Claude’s support for defining artifacts very helpful, as that makes it easier to keep track of and manage my growing set of contextual documents.
And finally, a cautionary result from an 8 month study at an 800 person tech company. This company did NOT mandate the use of AI. But as the study found,
AI tools didn’t reduce work, they consistently intensified it. […], we found that employees worked at a faster pace, took on a broader scope of tasks, and extended work into more hours of the day, often without being asked to do so.
Which seems like a dream come true for the company. But then, importantly…
That workload creep can in turn lead to cognitive fatigue, burnout, and weakened decision-making.
The report identified a few key reasons for why this was happening, including task expansion and more multi-tasking. Both of which I’ve definitely experienced over the past year. I can do so much more when I’m leveraging AI, but it also feels like I’ve got 20 spinning plates, and I have to keep them all going or I’m somehow failing.
I’d recommend reading the entire report (it’s not long). They have a few suggestions for how to mitigate these issues, though I have doubts about how effectively they will be.