Nathan,

A good "colloquial definition" depends on what you try and do with the
definition. If the intent is to clarify the potential role of AI in K-12
education, it is important to note that AIED has only recently started
to focus on data-driven AI, and many influential and commercial AIED
systems are knowledge-based.

In an attempt to "simplify" AI, it is easy to reify it and "clean up"
the complex history and reality of the field, as Pat points out.

We tried to address this in our recent European Journal of Education
article, Holmes & Tuomi: The State of the Art and Practice in AI in
Education. This is the abstract:

> Recent developments in Artificial Intelligence (AI) have generated
> great expectations for the future impact of AI in education and
> learning (AIED). Often these expectations have been based on
> misunderstanding current technical possibilities, lack of knowledge
> about state-of-the-art AI in education, and exceedingly narrow views
> on the functions of education in society. In this article, we provide
> a review of existing AI systems in education and their pedagogic and
> educational assumptions. We develop a typology of AIED systems and
> describe different ways of using AI in education and learning, show
> how these are grounded in different interpretations of what AI and
> education is or could be, and discuss some potential roadblocks on the
> AIED highway.
https://onlinelibrary.wiley.com/doi/full/10.1111/ejed.12533

This is available open access, with a number of other related articles
in the same special issue (The Futures of AI in Education).

Best,

ilkka

On 11.11.2022 10.50, Pat Langley wrote:
> Nathan,
>
>> Admittedly, our abstract uses a colloquial definition of "AI" as the
>> intended audience is K-12 educators who may not be technical experts
>> in the field. The nuance you helpfully described would be difficult to
>> fit into a 150 word abstract (and might even confuse the non-technical
>> reader). Hence, we provide the following clarification early in the
>> introduction:
> Thanks for the rapid response. The concerns I raised had nothing to
> do with "nuance", but rather with an unreasonably narrow definition
> of AI in the abstract that will reinforce popular myths.
>
>> "In general terms, AI refers to making computers do processes that
>> humans normally do. Similarly, ML is a major subset of AI that
>> involves computers learning from data and experience to perform a
>> task. Furthermore, DL is a subset of ML that uses artificial neural
>> networks, which are currently the mainstream of AI research (1,2).
>> In this paper, we loosely refer to AI, ML, and DL as AI although it is
>> important to know their nuances."
> I'm glad that you expand the definition of AI in the introduction and
> I hope you'll be able to do that in the abstract, too. But I would
> question whether deep learning is currently the "mainstream".
>
> Also, I encourage you to remove needless abbreviations. AI has been
> established since the 1960s, but "ML" is recent and "DL" is confusing
> because it also refers to the older area of "description logics".
>
>> With regards to the scope of this article, we argue that teaching AI
>> principles (particularly the supervised/self-supervised variety) can
>> be creatively rewarding for students and educators. Indeed, we focus
>> on machine/deep learning (as we make clear in the article) since
>> many practical applications benefit from these tools, which is central
>> to our argument.
> Machine learning for classification is arguably the least exciting area
> of AI research. Other branches focus on abilities that distinguish
> humans from dogs and cats, such as problem solving and reasoning.
>
> It's true that many successful applications focus on classification tasks.
> But that also held for early applications of machine learning and even
> expert systems, which implies that you should cover them as well.
>
> Best, -Pat
> To unsubscribe from this list please go to https://lists.aaai.org/confirm/?u=PgJygJLq5FbiWDRgozCDZUwXSH8qHw84