If AI Can Watch My Groceries, Why Can’t It Help Moderate Nextdoor?

One of Nirav Tolia’s favorite topics on the speaking circuit is AI.

That got me thinking.

If AI is such a core part of Nextdoor’s strategy, why isn’t it being used more effectively to improve moderation?

AI isn’t new. I’ve watched it evolve over the past 25 years, and today it’s part of my everyday life.

Take a trip to my local Harris Teeter.

I walk through self-checkout, scan my groceries, bag my items, and leave. Cameras, sensors, and AI are constantly evaluating what’s happening. When something falls outside an expected pattern or decision matrix, the system alerts a human associate to step in.

It’s AI first.

Human review second.

That’s a scalable model.

Now compare that to what I continue to observe on Nextdoor.

The attached example contains snarky comments that remained visible two days after they were posted. This isn’t an isolated example; it’s part of a pattern I’ve documented during my ongoing Nextdoor experiment.

Why isn’t AI identifying conversations that are escalating into personal attacks or unconstructive exchanges and routing them to trained reviewers?

Instead, Nextdoor continues to rely heavily on a decentralized network of unpaid moderators. While many volunteer with good intentions, any moderation system benefits from consistent standards, quality assurance, and ongoing coaching.

To me, the current model feels like the inmates running the prison while the warden sits in the office, removed from the chaos.

If AI can help prevent mistakes at a grocery store checkout, surely it can help create a more consistent and constructive online community.

Join the discussion on NielFlamm.com.