Howdy Helen, thx for bringing up a really important point about what
"natural" means in the context of human-computer interaction + who makes
the decision? Billionaire AI-Dev-Gatekeepers [who are clearly not equipped
(ethically or morally) for the job]? it also seems like OpenAI is quite
strategic [almost manic] about steering the mainstream AI narrative. They
released GPT-4o just a day before Google's Project Astra announcement:
possibly to preempt Google's news + discourse-dominate: talk abt an
aggressive approach to maintaining a lead in the AI landscape!

Your concern about what constitutes "more natural" interactions is spot-on,
ie the recent controversy involving OpenAI and Scarlett Johansson.
Johansson criticized OpenAI for creating a chatbot [named "Sky"] with a
voice eerily similar to hers without permission, highlighting how these AI
interactions are far from "natural" [granted, the term "natural" is
extremely problematic in itself + today seems so heavily weaponised
generally]. OpenAI ended up removing the Sky voice after Johansson released
a public statement abt it, but too late too soon IMO.

On Fri, May 17, 2024 at 6:24 PM Helen Varley Jamieson <
[email protected]> wrote:

> thanks for this, mez, really interesting. i find the claim of "more
> natural human-computer interaction" problematic - who is deciding what's
> "more natural" in this scenario? is it really "more natural" to need a
> machine to remind us to call our dad? that's an example in the promo video
> on the website of how useful chatgpt4 can be. & one of the presenters lists
> all the different controls they have developed to make ChatGPT 4 more
> appropriate/acceptable for society (annoyingly, the video is embedded in
> the website without any controls accessible, so i can't easily go back to
> the point to find her exact words). there is a whole ideology and agenda
> behind the software that needs to be deeply critically examined.
>
-- 
| mezbreezedesign.com
_______________________________________________
NetBehaviour mailing list
[email protected]
https://lists.netbehaviour.org/mailman/listinfo/netbehaviour

Reply via email to