[Cryptography] People vs AI

[Agathos]

On Sat, Mar 01, 2025 at 11:00:17PM +0100, Marek Tichy wrote:
> I'm also trying to find someone who would join me in pursuing the idea of
> having global autonomous digital identities anchored in a web of trust kept
> in a permissionless distributed ledger.

I have been working on this problem for over a decade.

Repo
https://codeberg.org/agathos/Eudaemon/
Whitepaper
https://codeberg.org/agathos/Eudaemon/src/branch/main/whitepaper.pdf
Release v1 POC
https://codeberg.org/agathos/Eudaemon/src/branch/main/releases

Proof of Human is the cornerstone of the networks security.

From the whitepaper published two years ago:

## 9. In defense against malicious AI

We hope that AI will be symbiotic with humankind.
However we are already entering an age where it is becoming increasingly difficult to ascertain the humanness of online content.
It is also becoming increasing difficult to ensure that those we are communicating with are indeed who we think they are.

Human powered sybil attacks are used to sway opinions and falsely represent the actual amount of people behind certain opinions and beliefs.
A single human can meaningfully manage perhaps at maximum 50 accounts when they are manually controlling them.
An AI powered sybil attack of the same nature has no real upper limit to the amount of fake accounts they can control.

If AI really gets out of hand then our greatest defense will be a small world web network of trust with all participants using strong asymmetric keys.


Agathos

-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 358 bytes
Desc: not available
URL: <https://www.metzdowd.com/pipermail/cryptography/attachments/20250310/38a4f177/attachment.sig>