Earlier quoted context omitted.
It used the words "freeze your hands"
I think any reasonable person would come to the conclusion it meant for you to chill your hands in ice water.
This post is about using AI to teach 5 year olds...
Building a real-time AI tutor for 5-year-olds
411–420 of 460 posts
Re: Building a real-time AI tutor for 5-year-olds
#412There is a privacy popup setting for: Marketing - These cookies are used to deliver relevant advertisements and track their effectiveness. Where do advertisements fit into the website and product? Ello has specific instructions to talk to the child about the learning activities. Giving large language models "specific instructions" is not a robust way to ensure safety. There really needs to be more published technical…
Hey there, I'm Tom - cofounder of Ello.
On the first Q: that popup is for the website. We're a consumer company and use paid marketing to reach parents. There are no advertisements in the product, nor is any data in the product sold or used for any purpose apart from delivering the experience.
On the second, we absolutely do not just tell an LLM to be safe and hope for the best. Every output of the LLM that is going to be presented to a child runs through a three-layer safety system. But - as importantly - everything a child says to the LLM goes through that same system.
1. Programmatic, anything that could easily be flagged with hardcoded rules. 2. An LLM evaluator that can interpret meaning to catch anything that the programmatic rules miss. 3. A human in the loop that reviews everything that could be slightly concerning.
We then, in turn, have a multi-tiered response system, including giving parents full access to everything that's happening so they're informed, flagging in a clear, but non-alarmist way, anything somewhat concerning, and finally direct outreach to the parents and/or authorities for anything critical.
Re: Building a real-time AI tutor for 5-year-olds
#413I think teaching a child to trust an LLM from a formative age is horrifically irresponsible. If anything, an app should be made where a child learns to correct an LLM's mistakes and learn that it isn't trustworthy. Actually, better, don't put an LLM in front of children. At all. EDIT: If a use case is for children who can't afford good education, then use an LLM to make educational materials for children, review them…
I'm Tom, cofounder of Ello.
-> I think teaching a child to trust an LLM from a formative age is horrifically irresponsible.
I largely agree, although with some nuance. One of the most important skills a young child today is going to need as they grow up is the ability to understand what's AI and what isn't, and how to interact with things that are. We actually work hard to make sure a child knows they aren't talking to something real - that they're talking to a computer program.
If part of the comment was on the accuracy of LLMs; our system of multiple harnessed LLMs running in parallel does deliver pedagogically solid and accurate content and experiences. There's a few comments in this thread about LLM hallucinations - We have curriculum specialists constantly evaluating the interactions that Ello delivers, and it just isn't a major factor right now.
So we need to ensure that we're delivering a high-quality, accurate experience for a child, while at the same time making sure that they understand what's reality.
-> Actually, better, don't put an LLM in front of children. At all.
I've responded above to your reason against using AI in education. But to judge whether it's right to avoid using AI with children entirely, we also need to weigh the positive side. So many children get a substandard education that isn't responsive to them and doesn't set them up for success - whether that's a child stuck in a class of 30 being treated as average, or one of the 270 million kids who don't have access to any education at all. Saying "don't put an LLM in front of children" misses the reality of what education actually looks like for so many of them IMO. If your kid gets everything they need from their school, their home, their private tutor and so on, then great.
So our answer is to use the benefits of AI to build technology that teaches - that understands where a child is and gives them the next learning experience they need to keep growing - while making sure they understand they're dealing with a computer program, and without trying to replace the critical role of the people in their lives.
Re: Building a real-time AI tutor for 5-year-olds
#414I think teaching a child to trust an LLM from a formative age is horrifically irresponsible. If anything, an app should be made where a child learns to correct an LLM's mistakes and learn that it isn't trustworthy. Actually, better, don't put an LLM in front of children. At all. EDIT: If a use case is for children who can't afford good education, then use an LLM to make educational materials for children, review them…
I'm Tom, cofounder of Ello. -> I think teaching a child to trust an LLM from a formative age is horrifically irresponsible. I largely agree, although with some nuance. One of the most important skills a young child today is going to need as they grow up is the ability to understand what's AI and what isn't, and how to interact with things that are. We actually work hard to make sure a child knows they aren't talking…
> We actually work hard to make sure a child knows they aren't talking to something real - that they're talking to a computer program.
The problem is not talking to a computer program. Kids have been using computers in school long before LLMs.
The problem is teaching a child to trust what the non-deterministic system is guessing at them. That will carry forward in their life.
You’re teaching them that an LLM is right, not just what it is.
As an aside, I’m also not sure about the certainty of what “making sure” a 4-year old knows is real or not is.
> If your kid gets everything they need from their school, their home, their private tutor and so on, then great.
That’s a powerful emotional point made to shut down criticism. Except it has nothing to do with putting a kid in front of LLMs.
I expressly mentioned this in my comment. You could create a mountain of educational material, apps, and tools with LLM assistance and verify it. Same result.
Re: Building a real-time AI tutor for 5-year-olds
#415Earlier quoted context omitted.
I'm Tom, cofounder of Ello. -> I think teaching a child to trust an LLM from a formative age is horrifically irresponsible. I largely agree, although with some nuance. One of the most important skills a young child today is going to need as they grow up is the ability to understand what's AI and what isn't, and how to interact with things that are. We actually work hard to make sure a child knows they aren't talking…
> We actually work hard to make sure a child knows they aren't talking to something real - that they're talking to a computer program. The problem is not talking to a computer program. Kids have been using computers in school long before LLMs. The problem is teaching a child to trust what the non-deterministic system is guessing at them. That will carry forward in their life. You’re teaching them that an LLM is right…
-> You could create a mountain of educational material, apps, and tools with LLM assistance and verify it. Same result.
I very strongly disagree with this. Teaching is not the same as a bunch of educational material and apps. The use of AI in education does not replace teachers, but it can get much closer to the responsive, learner centered, high-touch learning experiences that teachers provide, and that are proven to deliver the best outcomes.
-> That’s a powerful emotional point made to shut down criticism. You made a blanket statement - without any substantive points - that we shouldn't put LLMs in front of kids. I was arguing that there are advantages that we should take into account, not trying to shut down criticism.
-> not sure about the certainty of what “making sure” a 4-year old knows is real or not is. I agree. We're being very cautious here. My co-founder, who is a clinical child psychologist, has a blog post on exactly this you can find it in the footer of our website.)
Re: Building a real-time AI tutor for 5-year-olds
#416Earlier quoted context omitted.
The explanation I've heard is that the national curriculum switched away from phonics and to a teaching method called "balanced literacy", and this went terribly [1]. IIRC it involved teaching kids to recognize words by their overall visual shape. I believe this is what mature readers do, but the problem is that for this to work, one first needs to have bootstrapped a robust vocabulary. And that happens via being abl…
I've heard a lot of people argue that phonics are vastly superior to "whole word" techniques, and maybe that's true -- I'm definitely not an expert, though it is how I was taught English in Australia ~30 years ago. However, I find it quite hard to believe that it is the most important cause of the modern literacy rate issues in the US. Why? Because "whole word" teaching was the conventional wisdom since at least the…
Fascinating. Do you have any suspicions as to what the cause might be instead?
Re: Building a real-time AI tutor for 5-year-olds
#417Earlier quoted context omitted.
> In practice, yes. In theory a machines can do about everything a human can, but better and cheaper. If you could time-travel back to your 5y old self, would you prefer to be taught by AI tutor given the current state o/t art, or taught by whatever teachers you did have when you were 5? (with all the existing hallucination, breaking through guardrails etc problems of current AI in mind) If you'd have a ~5y old yours…
On the place of schools and peer-to-peer interaction: My oldest is about to start kindergarten in a few weeks. From what I can gather, she's already reading at approximately a mid-2nd grade level and doing math at a late-1st grade level. I expect that divergence would only grow if I kept her at home. So I already firmly believe my kids would learn much faster at home, and this is with us sporadically spending maybe 1…
One could look deeply into the Sudbury School or other similar approaches, especially for students with an existing internal drive to learn, as your daughter has demonstrated. The socialization aspects, particularly in participation in building group consensus in early developmental periods has far reaching effects.
If you believe that the core learning; reading, writing, and math competencies are being achieved at home, then a 'scheduled, unstructured time with peers and friends' may also be possible at this age through 'Head Start' program if available in your area.
The outcomes of Head Start's programs have been longitudinally validated from pre-kinder to the measurable outcomes through university graduation for a significant number of students across underprivileged and socioeconomically advantaged cohorts, with the process and results of these as a matter of public record by external researchers with independent review. Serious consideration may weight these types of studies more than "our students perform well" statements by other school projects.
That is not to say that other programs don't have merit, and in fact to the extent that the applied practices of any school overlaps with those studies, the results should be analogous.
I only bring this up, as at least for early grades, the 'hub for families' sounded like something that either the Sudbury or Head Start could deliver.
Disclaimers:
I have a family member who participated in every stage at Head Start; from in-classroom teacher, to multiple graduate degrees in developmental psychology and social work in order to design curricula, and as a regional manager for both HeadStart and a private community family social support caseworker. So I have some big picture concepts of what the program is and what it can provide.
I also have a close friends who have a daughter who was homeschooled through 3rd grade. Both parents have degrees in education and have been primary school teachers, one with a focus on linguistics and social studies, the other in math and life sciences. It really was Home+School. She then attended a traditional curriculum charter school, and then in middle school attended a Susbury school. The 'transition period' for her unfortunately was interrupted during COVID's early days. After the return to in-class attendance at the Sudbury School she staew there for a year and felt that she would perform better in a traditional classroom. Her parent's have been very complimentary of Sudbury approach, and they all feel that if she had been in that school from an earlier point in her education it could have worked for her.
It is clear that you are thinking deeply about this, and I wish you all the best of luck.
Re: Building a real-time AI tutor for 5-year-olds
#418Re: Building a real-time AI tutor for 5-year-olds
#419Earlier quoted context omitted.
The common sense part is not holding LLMs to a standard of absolute perfection that literally no other medium, or humans, can reach. You’re welcome to your views, but they don’t make a lot of sense, or fit with reality. Continue angrily tilting against windmills I guess.
You were skeptical that LLMs would hallucinate on topics for children and you said it was common sense. But now you’re saying the common sense is that they will hallucinate.
Man, it must be exhausting to have reading comprehension this bad.
Let’s try again: your claim is that LLMs “make up stuff all the time”, so they can’t be used to educate children.
I’m saying that modern SOTA models do not make up stuff all the time as it pertains to the topics a child of 4-9 would learn. That doesn’t mean they’re perfect or will never make mistakes, but there’s zero reason to think such issues would be widespread. And common sense would tell you that perfection is not required for LLMS to be a net positive in education.
I’m sure you’ll have another vapid rejoinder, but this has been boring for awhile and you seem to be enjoying being obtuse, so I’m moving on.
Re: Building a real-time AI tutor for 5-year-olds
#420Earlier quoted context omitted.
> We actually work hard to make sure a child knows they aren't talking to something real - that they're talking to a computer program. The problem is not talking to a computer program. Kids have been using computers in school long before LLMs. The problem is teaching a child to trust what the non-deterministic system is guessing at them. That will carry forward in their life. You’re teaching them that an LLM is right…
-> You could create a mountain of educational material, apps, and tools with LLM assistance and verify it. Same result. I very strongly disagree with this. Teaching is not the same as a bunch of educational material and apps. The use of AI in education does not replace teachers, but it can get much closer to the responsive, learner centered, high-touch learning experiences that teachers provide, and that are proven t…
We’re going to disagree about the quality of LLM learning experiences.
As an educator I’ve used them a lot for myself and in teaching others. Good for one-off topics if you check facts and a lot of human educators have built good quality content that it rips off.
LLMs are rather terrible at an extended progressive curriculum.
That said, the core thesis was simple: it’s dangerous to teach children to trust LLMs.
Not what they are. That they’re right. Which they aren’t. A lot.
You said you largely agreed with that. So, what’s the approach within this service to mitigate that?
I appreciate you responding but you’re throwing a lot of “but X” when the question is “what about Y”.