[ExI] AI can't invent (Legally!)

[BillK]

On Tue, 25 Aug 2026 at 18:19, Alex Blainey via extropy-chat
<extropy-chat at lists.extropy.org> wrote:
> The whole issue of AI, personhood, IP and liability is sticky and something Ive been wrestling with recently.
> Legal wrappers and incorporation can get around a lot of it and incorporation especially as it demonstrates how a non living entity (a company) can be given rights and legal status.
> It can be obfuscated by using holding and shell companies.
> I actually explored this idea for bootstrapping income for my Concord project. A fully automated AI run company creating digital products/services, which was itself subsidiary of a holding company.
> On the face of it totally legal and no one would know unless face to face was needed for meetings/signatures.
> _______________________________________________


The courts are still wrestling with this problem.
It will get more complicated as AIs become more independent, even self-aware.
The other thread goes into this in more detail.
But at present, when an AI misbehaves, the courts need someone to
blame. Either the AI developers or the user who set the AI off.

BillK