Human-First AI Pulse

Across six health systems, 263 clinicians let an ambient AI scribe draft their visit notes. Thirty days later, burnout had dropped from 51.9% to 38.8%, the mental load of note-writing was measurably lighter, and after-hours charting shrank by nearly an hour (Olson et al., 2025). The AI never touched the medicine. It took the paperwork, the part of the job that was quietly draining the people trained to heal. Here's the human read: the win didn't come from adopting more AI. It came from pointing one tool at the one task stealing the most energy. Before you ask what's trending, ask what's draining you.

Four cognitive scientists reviewed the evidence on what happens when we offload our thinking to AI (Cash et al., 2026). Their answer is calmer than the headlines: AI isn't making anyone stupid. The risk is quieter than that. When you hand a skill to AI, the practice that built the skill stops, and skills fade without repetition. Learners feel it most, but experts aren't exempt. For a credentialed expert, that's not a productivity question. It's an identity question, because your expertise is the gold your business is built on. Choose on purpose which parts of your work you keep doing by hand. That's where the skill, and the self behind it, stays alive.

Researchers followed more than 2,000 adults across four countries for a full year, tracking what happens when people turn to AI chatbots for companionship (Folk & Dunn, 2026). The loop ran both ways: leaning on AI for connection predicted feeling lonelier months later, and feeling less connected predicted leaning on AI more. The study doesn't say talking to AI is wrong. It says AI companionship isn't a substitute for being known by another person. If you build alone, and most of my people do, this is your alignment check. Let AI carry the work. Keep people in the seat AI can't fill.