Artificial Intelligence in Education - What's your experience?

[Jamie_Evans]

I work in a K-12 School in Australia and have been tasked by my principal to research and prepare a business case for our school to make a leap into artificial intelligence in education.

 

I'm lucky to have a principal that's open to these sorts of ideas and willing to make the investment so I want to make the very most of this opportunity.

 

 

Responses to any of these three question (if relevant to you) would be much appreciated!

Are you using AI in your school?

Is there something up and coming that I should keep an eye on?

What's your experience with AI in education?

[TechMonkey]

I'll be interested in this thread but surely the process should be, we have need x, how do we fulfill x, AI can do that and works better/cheaper than alternatives. Not we want to use AI, what can we do.

[Jamie_Evans]

I'll be interested in this thread but surely the process should be, we have need x, how do we fulfill x, AI can do that and works better/cheaper than alternatives. Not we want to use AI, what can we do.

 

It's not my job to question the process however I do believe we fill that criteria

 

-Our need is to educate our students in the most effective way possible

-We fulfill this need by investigating all all our options, as the IT Manager I've been tasked with investigating technology related options

[elsiegee40]

It's not my job to question the process however I do believe we fill that criteria

 

Actually, it is your job to question the process. You don’t just spend a fortune on $ShinyTech because someone has decided they want it.

 

I am speaking both as a governor and a school NM here.

 

What problems does whoever has charged you with this think it will fix?

 

“Improving teaching and learning” is too vague. Your SLT should know where weakness in T&L are ... they should be on the SEF (School Evaluation Form) and addressed in the SIP (School Improvement Plan)

 

It’s the solutions to those weaknesses that need to be found which may or may not include an AI approach. And if the problem is budget, then the processes that the school uses need to be fully understood before a decision can be made on whether or not automation might fix it.

 

Start with the problem and find the solution. Don’t start with a solution and look for a problem it might fix.

 

Without challenge you end up with a cupboard full of unused $ShinyTech and a big hole in the budget.

Edited April 5, 2019 by elsiegee40

[TechMonkey]

It's not my job to question the process however I do believe we fill that criteria

 

-Our need is to educate our students in the most effective way possible

-We fulfill this need by investigating all all our options, as the IT Manager I've been tasked with investigating technology related options

 

See I'd beg to differ. My job is to support the school to implement reasonable and suitable technologies to educate the students, supporting the teaching and learning. We fulfill this need by creating a specification in collaboration with the staff and investigating reasonable options. That does include highlighting cool stuff I've seen presented, announced or demonstrated, to staff. But it then has to have a valid application within the school that will be adopted.

 

So I would ask what is the specification you are trying to fulfill?

If it is to use AI then that is the tail wagging the dog, especially as AI is such a flimsy term. Do you mean actual AI or machine learning or algorithms?

 

There is some interesting work going on with mixing it with VLE and curriculum resources, take a look at https://www.century.tech/. I've seen talk about it being added in to systems like TurnItIn to make them more robust. There was a presentation talking about using it with Tracking modules to get better predictions of where a child should be and help build strategies to help them get there or improve. So there is work going on but if I was to just introduce the Century Tech system with out buy in from the school and it being embedded within the curriculum it will just be wasted money. If I bring the "AI" tracking in but the school isn't behind it and aren't recording the correct or pertinent data it will spit out rubbish.

 

Add to that the whole realm of ethics and "should we be trusting machine learning implicitly?" it gets a bit messy. More and more the conclusion being reached is that Machine learning is great when helping and supporting a human expert but on its own it is not reliable enough to be let loose, especially when even the developers of these systems do not totally understand how the results are produced so can not justify the findings.

 

A good book to read is Hello World by Hannah Fry. It goes into really good real world cases of Machine Learning being used and the consequences, sometimes unexpected and surprising, of these systems.