back to reflections

A Firewall for Attention

Social feeds are built to take your attention and hold it. A conversational AI can sit between you and the feed and only let through what you actually asked for. A firewall, in the original sense.
Petko D. Petkovon a break from CISO duties, building cbk.ai

We showed up for good reasons once. We wanted to belong to something, and to learn from people we would never have met otherwise. Somewhere along the way the deal changed.

Every social platform is built to take your attention and keep it for as long as it can.

What the feed hands you now tends to be one of two things. Either it inflates the bubble you already live in, more of what you nodded along to yesterday, or it is spam engineered to grab you by the collar and not let go. Both are tuned for time on screen. Neither has much to do with belonging or learning anymore.

For a long stretch the technology was working against us on this. The feed got smarter about us faster than we got wiser about it.

Conversational AI could be part of the solution. When I want to know what happened today, I ask the personal assistant I run myself. It sits between me and the raw feed and I tell it what I am actually after. The toxic stuff stays on the far side. I pull what I want instead of being fed what someone else decided I should see. That is a firewall. Something hostile on one side, you on the other, and a thing in the middle that only passes through what you asked for.

I already do a clumsy version of this by hand. Most of my news comes from a handful of channels I trust, our own Slack chief among them. The discipline of that is real, and it is tiring to keep up. An AI in the loop makes the discipline cheap. It does the filtering so I do not have to keep my guard up every waking minute.

This is the AI promise I find exciting. We spend a lot of breath on agents booking flights and closing tickets. The smaller, more personal win is that the same technology can hand us back the thing the feed has spent fifteen years taking.

For once, though, the tools are on our side.

Personal note: I do realise that there are dangers in this. The AI could be biased, or it could be manipulated by the platforms. It is not a perfect solution, but it is a step in the right direction. Alternative filters do not exist. And yes, we need to be vigilant about how these tools are developed and used.