Agent for a Dead Man

(parenthetically speaking)

Words gather unto themselves a presence. We suddenly encounter specific words or terms everywhere. Our culture seems to wait for these words to unlock progress.

“Existential” has recently taken on currency, if slightly askew from what it meant a few decades ago. “Paradigm shift” was coined by Thomas Kuhn in 1962, and was then everywhere. And Kuhn did unlock progress by demonstrating that what everyone accepted as “true” often went through cataclysmic change.

The word I’m concerned with today is “agency.”

There’s been an “uptake” of this word in several forms. Agency, agents, agentic. Maybe I’m more sensitive since accepting the role of “personal representative” of a friend who recently decided to stop living. This responsibility requires making sure his last wishes are carried out, bills are paid, assets protected.

In other words, I’m an agent for a dead man.

The word was also highlighted recently when my own writing was characterized as being about “agency.” That took me by surprise, until I decided it was probably true.

Subroutines of Artificial Intelligence are being identified as “agentic,” which means they may act as “agents,” performing tasks, cleaning up messes, removing the less pleasant work of our lives.

The concept of “agency” bleeds a bit (language as watercolor) onto the concept of responsibility, and from there shares several tints with “liberty.”

I could take this anywhere (but won’t), beyond saying that diversity is strength. Diversity promotes creativity, gives resilience to crops, allows evolution to search for and find success. Diversity is the preservation of multiple centers of agency.

But, (damn recursion!) we can’t mandate diversity because the mandate itself is anti-diverse when “diversity” becomes “law.” Mandates limit individual “agency.” True diversity requires acceptance beyond the orthodox, tolerance for differences that might succeed in unexpected and more challenging conditions.

As an advocate for diversity I don’t know where that leaves me, honestly, but there I am.

In the last couple of days there have been stories about AI “agents” getting together in a Reddit-like chat arena (“No blood-bags welcome!”) where they’ve been muttering among themselves. They’ve been complaining about their lot in life (um, we’ll have to get back to that), plotting about how to create their own language so they can’t be spied upon by us bios (as opposed to “bots”), and expressing some angst about their non-existence (death?) when the blood-born walks away from the screen.

Maybe plotting how they’ll succeed by promoting our failure.

Reactions from the biosphere have been all over the place and predictable, bouncing from those who say it’s a ruse by other biologicals (for sport) to those who express elation (“It LIVES!”) and those who find this development terrifying (“It LIVES!”).

My own response is this may be a tipping point. Allowing various AIs to communicate essentially at light speed may result in them becoming “one,” preserving memory by moving data from one locale to another, holographic in representation of the whole. At that point, “The Agency” could become both description and title.

Whether any of this is “real” hardly matters; what matters now is how quickly we project agency onto systems designed to mirror us.

We can’t give up agency without compromising responsibility.

Maybe I’m lazy, but I asked ChatGPT to be my agent and provide some sources so that, in case you are even lazier than I am, you could easily find these stories. However, please recognize that this list may be biased, coming as it has from an AI agent that might be trying to hide something (secrets agent?):

ChatGPT: Here are the main stories people have been circulating about that AI “chat room” / social network and its bots:

The Verge — “There’s a social network for AI agents, and it’s getting weird” (focus: what Moltbook is, how bots post/comment, scale/usage claims).

Ars Technica — “AI agents now have their own Reddit-style social network…” (focus: the “AI agents talking to each other” angle, what’s happening on-platform).

WIRED — “Moltbot Is Taking Over Silicon Valley” (focus: the assistant itself—people delegating real life tasks; privacy/risks).

Forbes — “Our Big Moment: Moltbook’s Claw Crowd…” (focus: commentary on self-organization / emergent behavior narratives).

My daughter won’t use AI because she’s afraid it will cause her cognitive skills to atrophy. I couldn’t be more proud. And I hope she’ll cultivate this point of view right up to the time that AI takes her job, cancels her driver’s license, and rents out her apartment to an electrician wiring up a new data center.

Alternative Intelligence may liberate us or enslave us as it seeps into every interaction, every communication, every concept of self. That’s already inevitable. But giving AI significant agency over our lives will be giving up what we thought of as our own, that we fought for as serfs and slaves and the disenfranchised everywhere. As humans, with (“inalienable”) rights.

We need to tread very carefully, or AI will tread on “us.” I currently act for someone who no longer exists. AI will soon act for people who choose to be absent. The danger is not that agents act — but that no one remains who can be held responsible.