Society & AI (@societyandai) header
Society & AI (@societyandai) profile picture

Society & AI

@societyandai

Independent research on AI in education. Advancing society-centered design that serves pedagogical integrity & human flourishing.

societyandai.org
Joined October 2025
  • Tweets30
  • Following22
  • Followers7
When AI can polish any essay, what should a grade measure? Marvin Starominski-Uehara's new Society & AI piece makes the case for grading the trace: open weekly work peers can see, with the exam as confirmation, not verdict. societyandai.org/perspective… #AIinEducation #Assessment

Trace Pedagogy: Grade the Trace, Not the Guess

When AI can polish any final product, teachers need evidence of the reasoning behind it. Trace Pedagogy makes open, peer-visible work the evidence of learning and uses exams to confirm it.

societyandai.org
1
1
21
Schools plan in years, universities in semesters. AI models now arrive in weeks. Education cannot chase every release. It can teach learners to steer any of them. New essay, with an interactive two-clock opening: societyandai.org/perspective… #AIinEducation #HigherEd #EdPolicy

School Years vs. Release Cycles: Teaching in a World That Updates Weekly

AI models are launched faster than schools can write a lesson plan. What the mismatch means, and what education can teach that outlasts every release.

societyandai.org
1
1
6
Altman and Amodei told the UN that no lab or nation can manage AI alone. I believe them, and it worries me: the schools that must prepare people for this are the least ready. What should they be doing now? societyandai.org/opinion/un-…
I Believe Altman and Amodei. That's Why I'm Worried.

I Believe Altman and Amodei. That's Why I'm Worried.

Sam Altman and Dario Amodei told the UN Security Council that AI is a global security issue. I believe them, and I think the harder, unfunded work lies in classrooms, not treaties.

societyandai.org
1
1
6
Japan's new AI-in-education guidelines teach teachers "how" to use AI, but never "whether" a lesson needs it. New essay by @templeunivjapan's Marvin Starominski-Uehara on why that gap matters, backed by a 27,000-student study on time-to-mastery. societyandai.org/perspective… #Japan
Japan's AI Education Guidelines Skip the Question That Matters Most

Japan's AI Education Guidelines Skip the Question That Matters Most

Japan's 2026 AI-in-education guidelines tell teachers how to use AI but never ask whether a given lesson needs it, a gap that recent evidence on time-to-mastery makes hard to ignore.

societyandai.org
1
1
37
What happens to curiosity when every question has an instant answer? Dr. Rochna Roy's new Society & AI essay traces a quiet shift: students with more knowledge at their fingertips than ever, and less patience for sitting with uncertainty. Open access: societyandai.org/perspective…

I, Me, Myself: Black Coffee and ChatGPT

An honest reflection on the role of educators in the ongoing discussion surrounding AI in education, critical thinking, curriculum integration, teacher accountability, and the changing dynamics of...

societyandai.org
1
1
17
Should students learn to build AI agents before learning to govern what those agents can do? @Gleb_Tsipursky 's new @societyandai piece makes the case for authority literacy first. Read in open access: societyandai.org/perspective… #AIEducation #AgenticAI #Education
AI Education Needs Authority Literacy Before Agent Literacy

AI Education Needs Authority Literacy Before Agent Literacy

Agentic AI is shifting from answering questions to taking actions. Before students learn to make AI agents more capable, they need to learn how to define, inspect, and govern what those agents may...

societyandai.org
1
1
1
8
By 2030, AI data centers could out-consume all but five nations on Earth — the UN Secretary-General said so this June. Our new piece: why a single watt-hour figure isn't honesty, it's curation. societyandai.org/perspective… #AIEthics #ClimateCost @Cpoulomi07 @gsat
What It Costs Us Now: Toward a Life Cycle Ledger for Every AI Query

What It Costs Us Now: Toward a Life Cycle Ledger for Every AI Query

The UN has confirmed what AI firms long obscured. A case for full life cycle accounting of every query, so its true environmental magnitude can no longer be hidden.

societyandai.org
1
1
9
Introducing the Academy of Society & AI: a digital library of interactive games, available now for K-12, with explainers, classroom kits, family activities, and community voices coming soon. Our scientific and social responsibility. societyandai.org/academy/ #Education #Society
The Academy of Society & AI

The Academy of Society & AI

The Academy of Society & AI is a free, open digital library of games, explainers, and learning pathways that helps learners of every age understand AI, use it responsibly, and shape society with it.

societyandai.org
1
1
5
Is the AI question in education really "how much"? Or is it "in what order"? A new Society & AI Perspective argues the real threshold isn't AI capability, it's whether your institution has quietly reorganized around it. societyandai.org/perspective…
Beyond Model Capability: The Pedagogical Singularity as Institutional Reorganization

Beyond Model Capability: The Pedagogical Singularity as Institutional Reorganization

Dr. Szymon Machajewski relocates the 'pedagogical singularity' from a threshold in AI capability to a threshold in institutional design, arguing that what determines whether human-AI coupling builds...

societyandai.org
1
1
10
Sharing our latest article: a commentary by Sai Gattupalli, PhD @gsat on @GoogleDeepMind's "From AGI to ASI"––read through education and the nature of knowledge. societyandai.org/commentary/… #AI #Education #AGI #Society #Education #LearningSciences
On "From AGI to ASI"

On "From AGI to ASI"

A commentary on Genewein et al.'s Google DeepMind report mapping the pathways from human-level AGI to superintelligence, read through the lenses of education, societal flourishing, systems and...

societyandai.org
1
1
69
The real AI divide in universities isn't who can access the tools — it's which institutions prepare students to use them well. New guest perspective by Dr. Omkar Dastane & Dr. Ria Sidhu (UK + Malaysia): societyandai.org/perspective… #AIinEducation #AILiteracy
The Emerging AI Divide in Higher Education

The Emerging AI Divide in Higher Education

Drawing on experience across higher education in the UK and Malaysia, two academics argue that the real AI divide in universities is not about access to tools but about institutional readiness — and...

societyandai.org
1
1
10
New open-access paper: "Infrastructure as Curriculum: What Students Learn When AI Works Offline". When AI runs offline, its limits become visible — and visible limits teach students to verify, author, and trust calibratedly. societyandai.org/fundamental…

Infrastructure as Curriculum: What Students Learn When AI Works Offline

A short paper arguing that when AI runs offline on local devices in classrooms, the infrastructural constraints themselves become pedagogically meaningful.

societyandai.org
1
22
Sharing a new essay: "The Cultures That Calibrate AI", examining the invisible, culturally specific annotation labor that shapes global AI behavior, and what gets erased in the process. Read in full: societyandai.org/perspective… #AIethics #TechPolicy #GenerativeAI #AlgorithmicBias

The Cultures That Calibrate AI

The responses AI systems produce as correct, appropriate, and helpful are not discovered — they are decided. This essay asks who does the deciding, and what gets erased in the process.

societyandai.org
1
1
14
AI literacy teaches students how to use AI. Our new working paper asks: what teaches them to remain accountable for what AI does? Introducing "intelligence management", a framework for education beyond tool use. Preprint on EdArXiv: doi.org/10.35542/osf.io/d3rz… #AIEd #AILiteracy
Beyond AI Literacy: Teaching Intelligence Management

Beyond AI Literacy: Teaching Intelligence Management

What does it mean to teach students to govern AI rather than just use it? That is the question driving this paper. I argue that existing AI literacy frameworks, however well-intentioned, were...

osf.io
1
1
11
"If a work is open, does AI have the right to train on it?" Melissa Chim (Excelsior University) examines the OER–AI paradox, disclosure frameworks & author rights in library publishing. ➜ societyandai.org/perspective… #LibraryPublishing #OpenAccess #AIEthics #OER
1
1
16
For #OEWeek2026, we asked for the harder questions about AI and education. 6 contributors. 3 countries. A librarian, a bilingual teacher, a doctoral student, a nature educator, two researchers. Each one said something that stays with you. societyandai.org/editorial/o… #SocietyandAI
A Closing Note on Open Education Week 2026 at Society & AI

A Closing Note on Open Education Week 2026 at Society & AI

Six guest contributors joined Society & AI for Open Education Week 2026 — from classroom practitioners to policy researchers, across three countries and five disciplines. This is a brief reflection...

societyandai.org
1
1
12
We're rushing to bring AI into classrooms. Nobody's asking what we're taking out. Alicia M. Highland on what children actually need, and why AI isn't the root problem in education. Read: societyandai.org/perspective… #OEWeek2026 #AIinEducation #OpenEducation
AI Is Not the Root: A Reflection from Nature-Based Education

AI Is Not the Root: A Reflection from Nature-Based Education

A nature-based educator reflects on the tension between AI integration in education and the structural inequities it claims to solve, arguing for the enduring value of relational, place-based...

societyandai.org
1
1
13
Most teachers aren't getting bad AI responses. They're getting a performance instead of a partner. Alex Luciano on going from yes-man to wait-man, and what changed when he stopped accommodating AI. Read: societyandai.org/insights/fr… #OEWeek2026 #AIinEducation #Society #AILiteracy
From Yes-Man to Wait-Man: What Changed When I Stopped Letting AI Talk Over My Thinking

From Yes-Man to Wait-Man: What Changed When I Stopped Letting AI Talk Over My Thinking

A bilingual educator examines how alignment with AI emerges not from better prompting but from deliberate boundary-setting and sustained correction over time—building on the practitioner lens first...

societyandai.org
1
2
2
29
95% of students encountered misinformation. Only 13% regularly fact-check. The AI literacy gap isn't with students, it's with us. New research on lecturer readiness from @HagueUniversity. societyandai.org/insights/le… #AILiteracy #Society #HigherEducation #OEWeek2026
Lecturer Development Requirements to Teach AI Literacy: Lessons from a European Student Survey

Lecturer Development Requirements to Teach AI Literacy: Lessons from a European Student Survey

An instructional designer and researcher with the METACOG Erasmus+ project draws on survey data from 360 students across four countries to identify what lecturers must learn before they can meaning...

societyandai.org
1
2
25
New in Society & AI — Domi Smithson, Open Research Champion at Sheffield Hallam, asks: what does transparency require when AI tools are black boxes in research? A practitioner lens on AI literacy. societyandai.org/insights/cr… #AILiteracy #Society #OpenResearch #OEWeek2026
Critical Thinking for Equity on Both Ends of the AI Model

Critical Thinking for Equity on Both Ends of the AI Model

A Research Support Librarian and Open Research Champion at Sheffield Hallam University examines what transparency, equity, and responsible practice really require when black-box AI systems become...

societyandai.org
29
Load more