AI used to be a tool on your desk. Now it is the desk. And the desk just hired itself as your assistant.
In our latest Active Teaching Lab, David Delgado shared how AI has become both the infrastructure for our work, and the experienced assistant able to anticipate our needs and execute complex jobs without being asked.
AI Baked in
AI is quickly transforming into an engine running the tools we use on a daily basis. It’s no longer about strapping AI onto an app. Apps are being built within AI itself. David shared how Adobe Firefly permits instructors to create consistent, editable images and infographics (all but impossible a year ago). Zoom’s AI Companion now summarizes meetings privately, with no data shared beyond participants. Microsoft Office 365 is likewise increasingly AI-driven: PowerPoint will create entire slide decks from text prompts, Excel will automatically analyze data, and Outlook can quickly summarize entire email threads. Finally, check your Canvas settings for experimental features like generated rubrics and discussion summaries. These Canvas features are worth exploring, though some may eventually move behind a paywall.
Agentic Labor as a Service
We are now entering an age of “Agentic Labor.” Tools like OpenClaw and Claude Code are capable of executing complex plans over hours and days rather than just generating single assets. In 2024, you might have asked an AI bot to draft an email to your supervisor asking for a raise. Today, you can ask AI to initiate a plan to get a raise. The agent will then evaluate your annual job reviews, compare your skills to the local and national job markets, examine the salaries for similar positions on job posting sites, and then walk you through a multistep process toward a raise, or even other qualified positions. In essence, AI has advanced from a robust version of Grammarly into an expert job coach who provides research, evaluation, and writing services.
This fundamental shift from AI producing “assets” to AI producing “labor” has real consequences for our students’ careers. Yes, AI can still transform raw data into a report in Excel. But employers will be expecting students to be able to manage a “fleet” of AI agents to craft and research business proposals, create large complex computer applications, or develop legal briefs. Thus we need to train our students as both subject-matter-experts and project managers who are able to employ AI agents and make quality assurance (QA) reviews of complex AI output. Content and project management expertise are the new essential hard skills for employability.
Check out our Podcast!
You may have heard of Notebook LM. It’s a Google-developed AI research and note-taking assistant. Notebook LM can analyze content you upload into the notebook (PDFs, Google docs, URLs, video) and use that to generate summaries, create mind maps, flash cards, answer questions, and yes, podcasts.
Following the February Active Teaching Lab session, I uploaded an edited transcript of the recording into Notebook LM and selected “Audio Overview.” Five minutes later I had a podcast file I then uploaded into Kaltura to generate closed-captioning. I used Gemini to transform the captioning into a transcript, and then uploaded everything to RSS.com. The entire process took approximately thirty minutes of actual work.
The result was a fifteen-minute conversation between two imaginary hosts who accurately broke down David’s main points and provided concrete examples to explain each one. There were a few inconsequential mistakes, but all-in-all the tool was more magic than math.
Check out the complete podcast at RSS.com, or simply search for “Active Teaching Lab” in Apple Podcasts and Spotify.