Matthew Karabinos

Drawing on earlier education-technology waves and social media, they show how adoption without purpose creates predictable problems.

• Intentional-design thinking supports deliberate limits and clear use cases.
• The position is a decision framework—screen time, purpose, implementation lessons—rather than a pro- or anti-AI stance.

In The Curve Ahead episode 6, Karabinos and Tracy Rains discuss the Los Angeles Unified School District's decision to limit student screen time. Drawing on the edtech boom, dopamine science, and Fred Rogers' philosophy of intentional design, they conclude that the answer to AI in education is neither fear nor full adoption, but balance.

The balance argument becomes more personal in the AI Real Talk episode: Karabinos credits Dr.

• Alfonso Mendoza Jr. with helping them pull back to become a “cautious advocate,” a term they attribute to their mutual connection Nneka J.
• McGee.
• That turns the public position into a personal caution about adoption.

Read original

Governments want reliability, auditability, and security assurances, while AI companies want to protect model integrity, safety constraints, and governance promises; those incentives do not always align. They frame the issue as a question of who gets access, under what conditions, and with what oversight.

Karabinos reports that the Pentagon is reconsidering parts of its relationship with Anthropic after the company declined to provide broad, unrestricted access to its AI systems, including Claude. The post contrasts that resistance with Anthropic's public reputation for safety documentation and asks what precedent the episode sets for the next generation of AI systems.

In the education conversation with Mark DiMauro, Karabinos lists DiMauro's governance points rather than Karabinos's own: DiMauro describes the governance problem as circular, identifies four competing definitions of the word “agent,” and points to an accreditation crisis. Karabinos presents these as DiMauro's framing, which connects the general governance argument to school-level questions about where guardrails should go.

Read original

Modeling questions, uncertainty, and learning in public builds student agency better than performing authority. Saying “I don't know” is a powerful teaching move, not a failure.

In the AI Real Talk episode with Stewart Brown, Director of Education Partnerships at Code for Kids, the conversation covers why teachers should not feel they need to be the expert in the room, what builds student agency in an AI-shaped world, and why saying “I don't know” can be a powerful teaching tool.

Read original · Original 2

Tools change faster than the concepts teachers need to explain, so foundational knowledge about how large language models and AI systems work should come before product training. This is a curriculum-coherence argument, not a claim about any single tool.

Karabinos asked six chatbot models—DeepSeek, ChatGPT, Gemini, Copilot, Claude, and Perplexity—the same prompt about the core concepts teachers should be able to explain about how LLMs and AI systems work.

• They report that most answers matched what they already cover in foundational sessions with schools and organizations.
• They then combined the responses in NotebookLM to produce an infographic and asked what teachers absolutely need to know is still missing.

Read original · Original 2

AI in education is unsettled enough that listening to teachers, administrators, and specialists across K–12, higher education, and corporate learning can produce better questions. The format also lets them test and revise their own stance in public.

Named shows include The Curve Ahead with Tracy Rains and AI Real Talk: The Curiosity Connection.

• Guests described in the posts include Mark DiMauro, Stewart Brown, and Dr.
• Alfonso Mendoza Jr.; episodes are promoted on LinkedIn with links in the comments.
• Karabinos also offers AI guidance, policy support, and professional development in the IU8 region and central Pennsylvania.

Read original · Original 2 · Original 3 · Original 4

Start here for a conversation that brings together recurring topics: classroom rules, governance, curiosity, and accreditation.

Matthew Karabinos talks with Mark DiMauro, a literature and computational linguistics professor, about how AI affects teaching.

• The post leads with DiMauro's line: “AI doesn't make you better at anything.
• It just makes you more prolific...”
• It lists the episode's themes: an Augmented Classroom Framework for consistent AI-use language on assignments, why the governance problem is circular, four competing definitions of “agent,” curiosity as a core student skill, and an accreditation crisis.
• Karabinos notes that DiMauro works across K-12, higher education, and corporate learning, and offers guidance, policy support, or professional development through Appalachia Intermediate Unit 08 in the IU8/central Pennsylvania region.

For brands planning a campaign

For an edtech professional development company proposing a K-12 AI literacy workshop series, this shows existing interview-based content and a regional audience. It supports a workshop module on classroom AI language, governance, and curiosity—not just tool training.

Start here for the signature podcast format and the balance-not-fear stance that recurs across the other posts.

In The Curve Ahead episode 6, Matthew Karabinos and Tracy Rains discuss what they call the inevitability of AI in education.

• Karabinos uses the Avengers and The Matrix to introduce the idea of inevitability, then connects it to AI moving through schools.
• The episode examines the Los Angeles Unified School District's decision to limit student screen time and asks what that signals about education technology.
• It draws on the edtech boom, social media, dopamine science, and Fred Rogers' philosophy of intentional design.
• They conclude that the answer to AI in education is neither fear nor full adoption, but balance.

For brands planning a campaign

This is a strong reference for a workshop series that needs a balanced-adoption framing rather than an either/or debate. It also suggests a possible podcast or co-host collaboration on screen-time policy and AI literacy.

Start here to see how Matthew Karabinos tests tools and turns model output into a teaching discussion.

Matthew Karabinos asked six chatbot models — Deepseek, ChatGPT, Gemini, Copilot, Claude, and Perplexity — the same prompt: what core concepts should teachers be able to explain and understand about how LLMs and AI systems work.

• The post reports that each model gave a large response and then condensed it into a one-sentence or framework summary, and that most of the information matched what Karabinos covers in foundational sessions with schools and organizations.
• They then copied the responses into NotebookLM and generated an infographic without further prompting, sharing the result and asking what teachers absolutely need to know about generative AI that is missing.

For brands planning a campaign

This supports a hands-on foundational workshop asset: a side-by-side model comparison, a teacher-facing concept checklist, or a co-developed AI literacy resource. Use it as a test and discussion starter, not as a finished curriculum.

Start here for the K-8 focus and the argument that computer science often exists in name only.

Matthew Karabinos interviews Stewart Brown, Director of Education Partnerships at Code for Kids, about why most schools are not ready for AI literacy.

• The post says the real gap in K-8 education is that “computer science” often exists in name only, scattered across coding clubs and robotics units without a foundation underneath.
• Their conversation covers why teachers should not feel they must be the expert in the room, what builds student agency in an AI-shaped world, and why saying “I don't know” can be a powerful teaching tool.
• Karabinos frames the episode as less about tools and more about the mindset shift schools need first.

For brands planning a campaign

For a K-12 AI literacy workshop series, this suggests an opening session on teacher mindset, student agency, and computer science foundations. Code for Kids is a plausible partner or comparison point for a school-facing program.

Start here for the “cautious advocate” idea that shapes how they talk about AI in education.

Matthew Karabinos introduces a new interview journey built around curiosity and asking questions.

• In this episode, they talk with Dr.
• Alfonso Mendoza Jr. about AI, the “tech-lash,” and lessons from classroom experience.
• Karabinos credits Mendoza with helping them see the bigger picture and pull back to become a “cautious advocate,” a term Karabinos attributes to their mutual connection Nneka J.
• McGee.
• The closing invitation is to get curious about AI and the future of education together.

For brands planning a campaign

This suits a professional development company seeking a thoughtful, caution-aware voice rather than a pure AI booster. It supports a workshop or event framed around curiosity, teacher experience, and responsible adoption.

Start here for the non-K-12 side of their AI commentary and their framing of institutional access questions.

Matthew Karabinos reports that the Pentagon is reconsidering parts of its relationship with Anthropic after the company declined to provide broad, unrestricted access to its AI systems, including Claude.

• The post explains that the U.S.
• Department of Defense has been evaluating AI vendors for national security uses and that some agencies sought deeper access for testing, auditing, and operational integration.
• Anthropic reportedly resisted open-ended access, and the Pentagon is said to be weighing supply chain or procurement consequences.
• Karabinos contrasts this with Anthropic's public safety and transparency reputation, then asks who gets access to AI systems, under what conditions, and with what oversight.

For brands planning a campaign

This supports a workshop series or commentary slot on AI governance, responsible access, and institutional policy. It shows how Karabinos connects AI safety questions to real procurement and oversight decisions.

The post says most responses match content covered in foundational sessions with schools and organizations, and asks what is missing from teacher-facing AI knowledge. This is a practical starting point for a workshop series on vocabulary, model behavior, and limits.

Multi-model prompting, synthesis, and visual summary; the post frames this as content tested in sessions with schools and organizations.

For brands planning a campaign

A K–12 AI literacy workshop series can open with a foundation module that explains core LLM concepts in teacher language and compares outputs from multiple tools. Turn the infographic and prompt approach into a participant activity rather than a slide deck.

Examples: Source 1

A post on the Pentagon and Anthropic raises questions about access, procurement consequences, safety commitments, and oversight when AI becomes foundational infrastructure.

• Another post with Mark DiMauro covers the circular governance problem, competing definitions of the word “agent,” and an accreditation crisis no one is talking about yet.
• The education focus is on where guardrails should go, not just on tools.

The post about Anthropic offers general AI governance commentary.

For brands planning a campaign

A workshop series can include a policy and governance module for administrators: acceptable use, assignment-level rules, oversight roles, and how to handle institutional exceptions. The national-security example can frame why access and audit questions matter, but adapt it to school policy rather than present it as a school case.

Examples: Source 1 · Source 2

The post references the ed-tech boom, social media, dopamine science, and Mr.

• Rogers’ intentional design.
• A second post with Dr.
• Alfonso Mendoza Jr. discusses the “tech-lash” and being a “cautious advocate.”

Podcast discussion and public commentary drawing on historical education technology examples.

For brands planning a campaign

A workshop can promote deliberate, balanced adoption by addressing screen time, purpose of use, and lessons from earlier education technology. This gives schools a decision framework instead of a pro- or anti-AI stance.

Examples: Source 1 · Source 2

Matthew Karabinos launched AI Real Talk: The Curiosity Connection and describes asking questions with favorite people.

• In one episode with Stewart Brown, the discussion covers why teachers should not feel they need to be “the expert in the room,” what builds student agency in an AI-shaped world, and why saying “I don’t know” can be a powerful teaching tool.
• Another post with Mark DiMauro calls curiosity the most important skill to build in students now.

For brands planning a campaign

A workshop series can train facilitation moves that reward questions, model uncertainty, and build student agency. An interview or guest-conversation format can show teachers how to explore AI without pretending to have all the answers.

Examples: Source 1 · Source 2 · Source 3

The conversation focuses on the mindset shift schools need before any of the tools matter.

Before commissioning a curriculum review, ask for evidence beyond this conversation.

For brands planning a campaign

An AI literacy workshop series for K–8 can pair AI content with foundational computer science concepts, teacher training, and a school-readiness check. This module targets curriculum coherence, distinct from policy or tool training.

Examples: Source 1

Named shows include The Curve Ahead with Tracy Rains and AI Real Talk: The Curiosity Connection.

• Guests in the posts include Mark DiMauro, Stewart Brown, and Dr.
• Alfonso Mendoza Jr.
• Episodes are promoted on LinkedIn with links in the comments.
• They also offer AI guidance, policy support, and professional development in the IU8 region and central PA.

Interview-based episodes distributed through LinkedIn and podcast platforms.

For brands planning a campaign

A brand can evaluate a co-produced interview series, sponsored episode, or guest-expert segment. The format suits campaign goals that depend on credible educator voices and discussion rather than product demonstration.

Examples: Source 1 · Source 2 · Source 3 · Source 4

The post speaks as “we” but does not specify their relationship with the organization.

View example

The posts do not specify the podcast’s ownership or sponsorship terms.

View example

Episode announcements frame the show as their journey.

View example

Dr.

• Jorge Valenzuela writes about project-based learning, assessment structures and visible thinking with a single-point rubric.
• Matthew Karabinos focuses on foundational AI literacy, governance and podcast conversations.

Read original · Original 2

Mary Harper's posts center on teacher-preparation design, the science of learning and early-childhood workforce policy in Texas. Matthew Karabinos focuses on AI literacy, AI governance and K-12 classroom practice.

Read original · Original 2

Dr Michael Harvey's posts focus on the shift from Google Gems to Skills and an educator innovation program. Matthew Karabinos covers broader AI literacy foundations, governance and podcast interviews.

Read original · Original 2

The posts argue that AI has made instructional design more important, and that project-based learning and strong assessment structures make student thinking visible: students justify decisions, explain reasoning, apply knowledge to authentic contexts, and reflect on process.

• This lets educators evaluate how AI was used rather than simply whether it was.
• The offer includes a single-point rubric a district can hand to teachers, and the questions address school and district leaders assessing the balance between AI literacy and instructional integrity.
• For a cohort whose deliverable is a rubric or a redesigned assessment, this is the closer fit.

The tradeoff is coverage of teacher-level foundations.

• Neither post says what teachers need to be able to explain about how large language models work, the opening module of the proposed Karabinos series.
• The longer instructional-innovation offers are described only briefly, so a buyer should confirm whether standalone AI literacy sessions exist before assuming the series is ready-made.
• Valenzuela's material reads as a case for a standpoint on assessment rather than a packaged multi-session workshop, while the Karabinos posts show the exact session activity a foundations module needs: one prompt run through multiple chatbots and compared for agreement and divergence.

Examples: Source 1 · Source 2

The posts centre on a design framework from the Learning Policy Institute that asks whether preparation systems are designed around what is known about learning and development.

• Its five principles cover coherence, clinical grounding, developmental science, equity, and modeling.
• The second post announces a Texas Workforce Commission grant to the Children's Learning Institute funding CDA credentials, registered apprenticeships, associate degrees, and bachelor's degrees.
• It targets Texas early childhood and EC-6, special education, and bilingual preparation.
• For a workshop series whose buyers are program deans or workforce-pipeline administrators, this writing is aimed directly at that room.

The tradeoff is distance from teaching AI in K-12 classrooms.

• Neither post discusses AI in K-12 classrooms, so the AI literacy content would need to be built from scratch rather than extended from demonstrated sessions.
• The posts present framework analysis and a grant announcement, without describing a workshop or professional-development format to build a delivery model on.
• If the need is to teach teachers how LLMs work and how to set assignment-level rules, the Karabinos posts already show that teaching, while Harper's material offers system-level design arguments for a program redesign committee.

Examples: Source 1 · Source 2

The posts take a position on who decides how AI enters learning: the change from Gems to Skills raises the question of whether teachers and students keep meaningful agency.

• They argue that teachers need to decide why, when, and how AI supports learning, while students need to decide when to seek help, question an answer, or reject it.
• The point is made concretely, with examples such as a teacher deliberately withholding a scaffold or requiring a prediction before an explanation.
• The Google Innovator posts show project work developed across three days with coaches and peer educators.
• For a district drafting rules about where judgement stays with people, these posts supply a sharper argument than a foundations session.

The tradeoff is breadth and demonstrated format.

• Two posts are available, both from late 2026, and both are argument and reflection rather than curriculum: neither shows a session plan, handout, or professional-development format, and neither covers foundational model literacy—what teachers should be able to explain about how LLMs work.
• A buyer who needs a three-session series with a concept checklist and an acceptable-use draft should ask Harvey whether they can develop it.
• The posts close on open questions rather than recommendations, which suits a policy discussion more than a hands-on teacher workshop.

Examples: Source 1 · Source 2