[ExI] Another ChatGPT session on qualia

[Adrian Tymes]

On Wed, Apr 26, 2023 at 1:10 PM efc--- via extropy-chat <
extropy-chat at lists.extropy.org> wrote:

> So basically, if the worlds opinion (philosophical, scientific or
> otherwise) starts to shift in favour of LLMs being conscious, this will be
> reflected in future training data, and the model might then,
> probabilistically start to say it is conscious?
>
> What I mean is that the machine will be conscious, when we as a species,
> via the training data "vote" that it is conscious through our tweets,
> scientific papers etc.?
>

Through changes in definition, perhaps.  If most people began to speak of
light at 440 nm wavelength as green instead of red, ChatGPT would reflect
this change too.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.extropy.org/pipermail/extropy-chat/attachments/20230426/227d0367/attachment.htm>