X-CGP-ClamAV-Result: CLEAN X-VirusScanner: Niversoft's CGPClamav Helper v1.23.2 (ClamAV engine v0.103.2) Return-Path: Sender: To: CNI-ANNOUNCE Date: Wed, 15 Mar 2023 15:15:31 -0400 Message-ID: X-Original-Return-Path: Received: from [73.160.73.173] (account clifford@cni.org HELO [192.168.1.32]) by cni.org (CommuniGate Pro SMTP 6.2.15) with ESMTPSA id 40421344 for cni-announce@cni.org; Wed, 15 Mar 2023 14:56:24 -0400 X-Original-Date: Wed, 15 Mar 2023 14:56:23 -0400 From: Cliff Lynch X-Original-To: cni-announce@cni.org X-Original-Message-ID: <20230315145623812119.024a6d00@cni.org> Subject: US Copyright Office Guidance on Works containing AI-generated material MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit X-Mailer: GyazMail version 1.7.1 There'a a very interesting document available from the US Copyright Office (it's due to be published in the Federal Register on March 16, 2023) that is intended to clarify the Copyright Offices position on registering works that contain material generated by AI. The document can be found at https://public-inspection.federalregister.gov/2023-05321.pdf As I read this (and I'm not a lawyer), the heart of the Copyright Office position is that only humans (not machines, not other living species) can produce works that may be registered for copyright, though of course those humans can employ various technologies to help in the creation of those works. The guidance then tries to explain what kind of help is permissible, and when creation of the work is essentially delegated to the AI technology, hence making the work ineligible for copyright. I will confess to a growing unease about the fundamental assumption here, though the implications of eliminating this assumption are also troublesome. My thanks, once again, to Gary Price for bringing this document to my attention. Clifford Lynch Director, CNI