AI Augments Human Judgment for Contextual, Sovereign Applications

Key Quotes

"And the obvious thing to do from a tech point of view is you build something that gives you the answer but that's wrong okay because humans are the ones that live in this world and we have to understand what's happening and when you build something it's it's like this is abstraction it sort of lacks the human element to it and also it's static so it's backward looking like all this ai stuff right it's all what people said five years ago you know 10 years ago so they can't keep up with current context they don't know how it feels to be a human so it's really important to have ai and humans work together where the humans can provide the context the perspective ai can help us and if you look at the places where we've been really really successful it helps us know what other people are thinking it helps us have conversations that are productive conversations and it helps us sort of reflect on ourselves to understand ourselves better and those are the three themes in the book and so that avoids the alignment by leaving the alignment with the people it's only when you try to replace people that you want to have things that are doing what we want and doing what we want well i don't know about you but i have a hard time knowing what i want"

Dr. Alex Pentland argues that building AI systems solely to provide answers is a flawed approach because it removes the essential human element and context. Pentland explains that AI, being static and backward-looking, cannot keep pace with current realities or understand human experience. He advocates for a collaborative model where humans provide context and perspective, enabling AI to assist more effectively in understanding others, fostering productive conversations, and promoting self-reflection.

"And the obvious thing to do from a tech point of view is you build something that gives you the answer but that's wrong okay because humans are the ones that live in this world and we have to understand what's happening and when you build something it's it's like this is abstraction it sort of lacks the human element to it and also it's static so it's backward looking like all this ai stuff right it's all what people said five years ago you know 10 years ago so they can't keep up with current context they don't know how it feels to be a human so it's really important to have ai and humans work together where the humans can provide the context the perspective ai can help us and if you look at the places where we've been really really successful it helps us know what other people are thinking it helps us have conversations that are productive conversations and it helps us sort of reflect on ourselves to understand ourselves better and those are the three themes in the book and so that avoids the alignment by leaving the alignment with the people it's only when you try to replace people that you want to have things that are doing what we want and doing what we want well i don't know about you but i have a hard time knowing what i want"

Dr. Alex Pentland contends that technology designed to simply provide answers is fundamentally flawed because it overlooks the crucial role of human understanding and context. Pentland elaborates that AI, by its nature, is static and based on past data, making it incapable of grasping current situations or the nuances of human experience. He proposes that AI should function in partnership with humans, with people supplying the necessary context and perspective, thereby enhancing AI's utility in comprehending others, facilitating meaningful dialogue, and encouraging personal insight.

"I mean i see ads i hear what happens on on you know streaming tv etc but it's usually only when you get into it that you realize hey this is not fun or wow this is a lot better than i thought it would be having things in line with that when we don't even know it seems a little bit unlikely if ai is ultimately training on human created material why isn't the output more human centric when all of the input is there's two things here one is you know what it's training on is just what people say on the internet it's not even what they say in private and it doesn't have all of the feeling and experience of actually being in a place so it's just saying oh this is what people say and you know you and i do this all the time also it's like if someone comes up and says hey what about x right you can give a paragraph of answer you may know almost nothing about it but you've heard things and you know that's what ais are they're sort of like the common conversation don't live your life according to what you hear you know in that sort of answer because some of it's right but some of it's not right and in particular it's not necessarily right for you"

Dr. Alex Pentland questions the human-centricity of AI output, noting that even with human-created input, AI often fails to align with individual desires or experiences. Pentland explains that AI trains on public internet data, which lacks the depth of private sentiment and lived experience, resulting in responses that reflect common discourse rather than personal truth. He likens AI to casual conversation, providing information that may be partially correct but not necessarily tailored or appropriate for the individual user.

"But now you're much more tuned in through this little uh it's like a smart librarian ai uh than you were before so you're part of the culture better you know what's going on you can be much more appropriate it's a really really sort of simple to do and b a little bit of genius you know technology and ai are both things that very much exist kind of without borders but as an advisor to the abu dhabi investment authority lab having worked with the eu and the un secretary general's office you know you've seen how different countries and different governments deal with the challenges and the opportunities that are posed by new technologies so the technology is somewhat without borders but what we do with it very much comes down to the governing body how are other countries addressing ai compared to the way the united states is are they doing something that we should be doing or is there something that we should all be doing together that we're really not yet yeah"

Dr. Alex Pentland highlights the potential of AI, exemplified by a "smart librarian" analogy, to enhance cultural awareness and appropriateness within organizations. Pentland suggests that such AI tools can make individuals more informed and integrated into their company culture, describing this as a simple yet ingenious application of technology. He then pivots to discuss how, despite technology's borderless nature, its implementation is heavily influenced by governing bodies, prompting a comparison of international AI strategies.

"And the other one to look at is loyalagents org just to sort of understand you know what it is that we'd like to have these ais do and what standards we ought to hold them to to be able to really say that they serve us and not just you know another persuasion engine so those are two things and you know my email is out there but i get a million emails a day and i'm sorry well professor patlin thank you so much for making the time uh we've really enjoyed this and we know our audience is going to get a lot out