Will ChatBots become an indispensable tool for most children's K-12 grade education in the United States before the end of the decade (<2030)?
➕
Plus
43
Ṁ3878
2029
50%
chance

most: 50% of students in the United States

indispensable: allowed and expected to be used in and outside of class; parents and teachers (not 100%, but many) would consider lack of access to be equivalently disadvantaging to not allowing students access to books or computers.

indispensable 2: it's not required that all activities permit the use of ChatBots, e.g. tests, just that overall they're considered indispensable.

Get
Ṁ1,000
and
S3.00
Sort by:

https://www.hackster.io/news/clever-student-has-ai-write-their-homework-by-hand-4087f25994c2

Forcing a student to write papers by hand may cease to be a speed bump.

predictedYES

@RealityQuotient It's not like most students have access to this kind of thing.

@osmarks True. If you don't believe it's likely that that will change by 2030, that should weight heavily on your wager.

predictedYES

@RealityQuotient I don't have hugely strong opinions on whether future children will have more engineering abilities, 3D printers will get cheaper, etc, but I don't think it matters much. Copying an essay off your computer by hand is still faster than writing it manually, and in any case the question's framing requires that chatbots be approved of by the education system, not just individual students.

No...ChatBots are powered by large language models. Large Language Models have been shown to be most stochastically approximate people like us, who are online all the time, who are mostly male, middle or higher income, and do not approximate individuals who don't contribute a lot of text online...meaning...minorities.

So yeah, a school district is not going to adopt that unless there is a massive amount of re-biasing to ensure that all students' needs and perspectives are met. You may get some adoption, but not mass adoption.

So we need some kind of ed-tech specific LLM. Complicating that, ChatBots are INCREDIBLY expensive to build and maintain, they are State of the Art. While that expense will no doubt go down, it's going to take another 3 to 5 years, or maybe 10 years to see what we get with ChatGPT, with instantaneous next word prediction being at the levell of affordability as say...something that a smaller company could run. EdTech does not tend to get huge, high-end investment, because it moves slow, because you have to sell to State Governments and individual school districts and schools, which is a slow process.

I think unless you see some dramatic drop in the price of GPUs, you're just not going to get things down to a price point where it's investment grade for EdTech companies to play around with and sell into school districts.

Schools and individual teachers will dabble around with ChatGPT, but to say that it's indispensable, that it's part of the curriculum, vs. there could eventually be a backlash, because of the biases inherent in the main ChatGPT..that makes it a huge unknown.

On the other hand, Microsoft could specifically try to go after the EdTech space and modify ChatGPT to fit their criteria.

predictedYES

On the other hand, Microsoft could specifically try to go after the EdTech space and modify ChatGPT to fit their criteria.

I feel like you're underestimating how doable this is in seven years. if this were about six months out I'd agree with your take.

Buying yes to sell after GPT-4 release hype and gwern analysis.

How would you define a ChatBot? Do you have in mind a broad definition, such that any kind of AI capable of ChatGPT-like chatting would qualify, or a more narrow definition? Is GPT2 a ChatBot? What about ChatBots embedded in other products, like YouChat?

@whenhaveiever I'm open to definitions.

That said, if people find it indispensable, it'll be clearly a few steps up from Eliza: https://en.wikipedia.org/wiki/ELIZA

predictedYES

@RealityQuotient what if it's a full strength agentic agi capable of embodied interaction through a robot body, or something similarly way beyond chatbot level?

@L I would still resolve YES. While it would obscure whether or not today's capabilities were sufficient, this market is more about whether the technology will be depended on than on what the technology is. If the capability of today's (ChatGPT) ChatBots or better are indispensable by a majority, I'll resolve the market yes.

50% seems like a high bar. What was the penetration rate of say, personal computers in schools? Probably took until the early-mid 2000s for >50% of most schools in the US to have computers? So that's a 10-20 year post-invention lag?

@jonsimon You're correct. The technology would have to be adopted very quickly to resolve YES.

@jonsimon Comparing the invention lag of no computers -> computers in schools is not fair because computers were solving the infrastructure problem in schools. The adoption of computers meant building infrastructure and redesign of curriculum, setting up of computer labs, etc. However, adoption of software moves in quicker cycles, because the infrastructure to run it on already exists (aka computers, smartphones, tablets - the bottom line is pretty cheap). Still, "most" K-12 schools sounds like too much.

© Manifold Markets, Inc.Terms + Mana-only TermsPrivacyRules