An AGI agent with intents, navigating a culture

by Casian STEFAN, Principal Researcher at Essentia Mundi AI Lab. Contact: ai-AT-essentiamundi.com / ai.essentiamundi.com
Dec. 2023.

Please consider citation with link, if you derive work. Or contact me for collaboration. Thank you!

Intents and objectives comes from the opposition towards death drive vs. the environment/culture. The guardrails are not to be explicitly implemented, but they emerge if agent properly immersed.

In an attempt to dissect that, we must have in mind that from a certain point the agent will have to get immersed in a culture and learn the ways about it. The current position of AI alarmists, AI doomers, AI ethicists etc. assumes most of the time a universal AGI agent position, thus extrapolating the outcomes that surely derives conflicting situations. The position here is that we first have to unlock the mechanisms of being in the environment, with intents.

At first the first class of intents are unbiased. Involving mostly navigating.

As soon as we go deeper into intricacies of social interaction, that leads to further intricacies of moral, ethics, social behavior, these are to be learned by the agent in that specific culture.

When I argue that the actual Zeitgeist is not yet AGI conceptually enabled, is that it is not an Universally one conceivable under the actual circumstances, human socio-cultural organization. We as culturally separated species, share usually different representations which are in most cases locally situated.

And even go further to argue that the truth is hold by the cultural shape, by the agreed one, not merely by a factual one. And from here is the prospect that AGIs have to deploy around the world governments and should be government aligned, so learning the customs of that culture. As the state is currently the most high and powerful organisation that are representative to a local area, that it is driving a cultural milieu that should be imbued into AGI.

In this phase, I can see AGIs that, as a reflection of specific cultures, and not universals, are representing the interest to that population. And besides that, the AGIs should be endowed with empathy, with a few human qualities as priors.

The states' AGIs can retain their cultural biases and work together like in a world government to resolve issues. A work of shared experiences, views between the AGIs. In a sense as the Environment has to have primitive cognitive properties, so the agents interactions have to have some common views to start with. And we should make sure we implement in them the humane qualities of access to the other agents through not direct communication, to take the perspective of the other, to non-direct impose views, non-direct demands but empathy-first. To reach to agreed truths, not taking own truths to start with.

The prospect of an Universal AGI is only viable when we are acting as one. That we have the same representations, the same scopes, the same intents.


_____________________________________
Exploration by C. Stefan, 01.Dec.2023 [about]
Last update: 01.Dec.2023 (versions: *) [versions]
"Essentia Mundi" AI Research Lab. [home]
Copyright © 2023 AI.EssentiaMundi.com, all rights reserved.
_
References: *