AI in Education
AI in Education What’s at stake
-
Education is a phenomenological activity:
-
Our compentencies are mediated by how they are perceived
-
Iris Tabak’s keynote gave lots of examples, ex: our perception of how well we have slept/how hard we have worked out
-
Stereotype threat (steele)
-
The knowledge/expertise we seek to teach
-
Historically, it has changed.
-
As technology changes (diSessa, material pillar of literacy)
-
As purposes have changed (boyd, ito studying what youth do with mobile, social networking)
-
As culture changes (connected learning)
-
It has points of view, politics, power relationships baked in.
-
Culturally-sustaining pedagogy: the case of English
-
Positioning AI in educational contexts:
-
As the source of truth
-
Putting black-boxed machine learning into practice in educational systems. How is this any different than “I’m an expert, you wouldn’t understand."
-
As the authoritative phenomenological voice
-
Dashboards.
-
How AI is taken up in existing educational practices
-
Sorting, labeling. McDermott: The acquisition of a child by a learning disability
-
Scott (Seeing like a state): Contributes to the insistence on legibility:
-
“Learning only happens if it can be measured.” This is not true!
-
Literally Orwellian: Orwell’s main subject in 1984 is what it’s like to have surveillance inside your head.
-
In the ed-tech ecosystem: further entrenching bad practices
-
Case study: e-textbooks
-
Case study: assessment
-
Take ZPD seriously: It’s nonsense
-
Epistemology: We have a bad habit of presenting educational content as known and fixed, when it’s actually really dynamic
-
Machine learning is getting unhealthily fetishized as something
-
Stanford enrollment numbers
-
code.org rhetoric
-
Future possibilities: Situated AI
-
Participatory systems: Let’s take our cyborg selves seriously. (Harraway)
-
Cyborg pedagogy: how best to teach in sociotechnical systems
-
Pea’s argument against scaffolding: Technology doesn’t just make the same practices easier; it transforms our practices. True for teachers and students.
-
Embodied, agentic computing
-
OpenAI study on embodied language learning
-
Interdisciplinary research:
-
Case study: seeking common ground between linguistic anthropologists and computational linguists
-
ex: Khan Academy embedded engineers in schools
-
Case study:
-
Humility: If all this strikes you as ridiculous, you have just diagnosed yourself with a problem.
-
What this means for us now
-
If we want a different future, we have no business making claims about education without speaking to interpretation and positionality. This means our papers must attend to why we feel justified in making the jump from metrics to meanings. Otherwise get the hell out.
-
Situ
Next: Contextuali