Re: Ansible 455 -- June 2025

On 5/30/25 11:54 AM, David Langford wrote:

RUMBLINGS. _Seattle Worldcon 2025:_ all fandom was plunged into war, or so
it seemed, by the admission on 30 April that Seattle had vetted potential
programme participants with ChatGPT (to the especial annoyance of creators
whose work was used without permission or payment to train this Large
Language Model). An apology from con chair Kathy Bond followed on 2 May,
and a much longer statement on 6 May. The latter revealed that rather than
asking whether applicants were good speakers or moderators, the prompt fed
to ChatGPT began: 'Using the list of names provided, please evaluate each
person for scandals. Scandals include but are not limited to homophobia,
transphobia, racism, harassment, sexual misconduct, sexism, fraud.' LLMs
were also mentioned, perhaps not incidentally, in a public resignation
statement by WSFS division head Cassidy and Hugo administrators Nicholas
Whyte and Esther MacCallum-Stewart (Bluesky, 5 May). In Bluesky comments,
Nicholas Whyte quasi-explained: 'Frankly one can get tired of fighting all
the bloody time.'

Considering that most of the terms in the list are routinely used to smear people, the chances of the query returning trustworthy results are extremely close to zero.
--
Gary McGath    http://www.mcgath.com