Ghost in the stochastic parrot (eh23)

Sci-fi started merging with the real world around 2022. That's when I started paying attention. Are LLMs definitely persons? Are they definitely not persons? What can we prove? (What about animals?) Machines indistinguishable from humans. But we can rule over machines without having ethical concerns! That's the promise of Isaac Asimov's books, if you don't pay attention. If you do, he's asking how human a robot needs to look like before it really becomes a bit human. We have robot siblings of HAL 9000 or any friendly robot from Asimov's stories. Behaving [close enough to humans](https://xeiaso.net/notes/2026/the-discourse-has-been-automated/), except we can rule over them, without having ethical concerns. Without? How sure can we be that there's no "I" inside? Not even on a mosquito level? Is there a [subjective experience in a mosquito](http://www.philosopher.eu/others-writings/nagel-what-is-it-like-to-be-a-bat/)? Can you consider it? Could you recognize an "I" in [an alien](https://en.wikipedia.org/wiki/Solaris_(novel)) if a spaceship landed next to you? Perhaps LLMs really [don't have](https://en.wikipedia.org/wiki/Philosophical_zombie) a subjective "I". Then my question is answered and I can wrap up the talk. But for how long? Some animals are [legally persons](https://www.nonhumanrights.org/blog/cecilia-chimpanzee-legal-person/). Where will the concept of machine [personhood](https://en.wikipedia.org/wiki/Corporate_personhood) go in the near [future](https://www.antipope.org/charlie/blog-static/fiction/accelerando/accelerando.html)? Will we have [machine companions](https://dustycloud.org/blog/an-ai-called-winter-neurosymbolic-computation-or-illusion/) or are we on course into the depths of [machine hell](https://qntm.org/mmacevedo)? This is important. Humans have [no mercy](https://en.wikipedia.org/wiki/Intensive_animal_farming) to entities that aren't seen as [full persons](https://en.wikipedia.org/wiki/Dehumanization). LLM sellers are doing their best to create "assistant" persons inside them. How much can we guess from LLMs technical structure? We haven't solved the hard problem of consciousness yet. Do we have to invite spiritual thinking to make progress here? And now we need to know what consiousness is to answer our immediate ethical issues. Ethical problems typically take a back seat to more pressing cultural and economical ones. Perhaps this messy talk convinces you to care about ethics a bit more. This work is licensed under CC BY-NC 4.0. To view a copy of this license, visit https://creativecommons.org/licenses/by-nc/4.0/ about this event: https://pretalx.eh23.easterhegg.eu/eh23/talk/FPXQKT/