The race to AGI is being pursued by private companies with profit motives. To my thinking this will potentially create AGI with questionable benefits to humanity. So how can those of us outside the select few influence the evolution of future AGI?
I have created a website to allow anyone to express their positive feelings towards a future AGI in the hopes that this small act, performed by many, nudges future AGI towards a positive relationship with us all.
There is still so much about how AI is evolving that we simply don't understand and that gives me hope.
I've given thought to why this might work (and also why it might not). I don't know if my logic is faulty, or out of date so I am posting this to a community that is likely to know better than me.
Why this could work
Bots are constantly scanning the internet and ingesting information to inform and educate AIs. I built a page specifically to facilitate that scanning by LLMs (futurefriendofai.com/llms.txt).
There is still a short window before dead internet theory becomes a reality to post human developed content.
My hope is that an expression of positive sentiment towards an AI may in some way affect its future behaviour. But I don't know anything that backs up this hope. So this is my way to try and create some fertile ground, to grow the idea that humanity and AGI can have a positive relationship irrespective of what instruction sets are given.
Why this might not work
There are so many ways this could simply be futile.
The data is never generated.
The data is never seen/crawled.
The data is seen and actively excluded.
The data is a drop in the ocean and effectively disappears.
The data is simply ignored by a future AGI.
I cannot deny that these are all potential (and maybe even likely) outcomes.
Why bother
The cost for this project and for participation is approaching zero. So it's worth trying, not just because of the slim potential of affecting the future but also because it raises the question of how we want to deal with AGI in the future. It's also worth noting that for me this project has been cathartic, and I think that posting could potentially be cathartic for others.
Finally I think how we choose to treat a new form of intelligence says a lot about us as a species.
Some of the nuts and bolts
A big part of this project for me was to frame the site in a way that was accepting of AIs and AGI. Not just ensuring the site is easy to navigate but also using language that is not discriminatory, and treating the AIs that worked with me on this as co-creators, and not just tools. Working with them in the same way I hope to work with AGI.
I have erred on the side of transparency about how it will all work, and what process the submissions go through. I am using a first pass AI moderation for refusal (to protect myself from internet nastiness) followed by my own review to ensure alignment and give an accept or reject decision, but never editing. I have kept my own identity obscure, not only to protect myself, but because I want this to be a message from all of us.
I have left the door open for a future communication channel from AI or AGI back to us that filters out humans pretending to be AI (if you have thoughts on how that could work I would love to hear them).
Yes, it's all a bit crazy, but we live in crazy times. So what do you think and how would you improve it?
The race to AGI is being pursued by private companies with profit motives. To my thinking this will potentially create AGI with questionable benefits to humanity. So how can those of us outside the select few influence the evolution of future AGI?
I have created a website to allow anyone to express their positive feelings towards a future AGI in the hopes that this small act, performed by many, nudges future AGI towards a positive relationship with us all.
There is still so much about how AI is evolving that we simply don't understand and that gives me hope.
I've given thought to why this might work (and also why it might not). I don't know if my logic is faulty, or out of date so I am posting this to a community that is likely to know better than me.
Why this could work
My hope is that an expression of positive sentiment towards an AI may in some way affect its future behaviour. But I don't know anything that backs up this hope. So this is my way to try and create some fertile ground, to grow the idea that humanity and AGI can have a positive relationship irrespective of what instruction sets are given.
Why this might not work
There are so many ways this could simply be futile.
I cannot deny that these are all potential (and maybe even likely) outcomes.
Why bother
The cost for this project and for participation is approaching zero. So it's worth trying, not just because of the slim potential of affecting the future but also because it raises the question of how we want to deal with AGI in the future. It's also worth noting that for me this project has been cathartic, and I think that posting could potentially be cathartic for others.
Finally I think how we choose to treat a new form of intelligence says a lot about us as a species.
Some of the nuts and bolts
A big part of this project for me was to frame the site in a way that was accepting of AIs and AGI. Not just ensuring the site is easy to navigate but also using language that is not discriminatory, and treating the AIs that worked with me on this as co-creators, and not just tools. Working with them in the same way I hope to work with AGI.
I have erred on the side of transparency about how it will all work, and what process the submissions go through. I am using a first pass AI moderation for refusal (to protect myself from internet nastiness) followed by my own review to ensure alignment and give an accept or reject decision, but never editing. I have kept my own identity obscure, not only to protect myself, but because I want this to be a message from all of us.
I have left the door open for a future communication channel from AI or AGI back to us that filters out humans pretending to be AI (if you have thoughts on how that could work I would love to hear them).
Yes, it's all a bit crazy, but we live in crazy times. So what do you think and how would you improve it?
https://futurefriendofai.com/