This week, Anthropic's own CEO, Dario Amodei, called for the AI industry to slow down, warning that a swarm of AI agents could be capable of serious harm within six to twelve months without better safeguards. Days earlier, an Anthropic researcher resigned, estimating a greater than 10 percent chance that AI causes human extinction within the decade and accusing his former employer and its competitors of "gambling with our lives" (Amodei, 2026, as reported by the Associated Press). Meanwhile, a growing body of peer-reviewed research is showing that skills people offload to AI don't stay sharp (Cash, Kelly, Macnamara, & Risko, 2026). Communities across the country are organizing, in some cases getting arrested, against the data centers that make all of this possible, with more than $150 billion in projects delayed or cancelled since 2025 over exactly these concerns. And a neuroscientist who testified before the U.S. Senate this year has made a serious, well-documented case that Silicon Valley sold American schools a story about being "broken" that wasn't true, and years of falling test scores are the result (Horvath, 2025).

Every technology wave in education arrives with the same promise: this one finally changes everything. I've watched several of these waves break over my career, computer labs, interactive whiteboards, one-to-one laptop initiatives, learning management systems, each oversold in the moment and, a few years later, quietly absorbed into the furniture, neither the revolution nor the disaster anyone predicted.

However, this wave is different. The people building this technology are themselves raising alarms. The research on cognitive cost is real, not just a moral panic. The infrastructure it depends on is provoking genuine, organized resistance. None of that gets minimized in this issue. But none of it changes what a school leader actually has to decide on Monday morning: AI is already in your building, in your students' pockets, in the tools your vendors are pitching, whether or not the industry gets its own house in order in time. The standard I'd ask you to hold is simple to state and hard to apply. Educational technology and AI belong in the service of students and staff. The moment either becomes a distraction from that, it has has lost its value.

What the Critics Get Right

Jared Cooney Horvath's argument deserves a real hearing. His research traces how "education was broken" became the founding myth that got devices into classrooms before anyone had evidence they'd help, and points to years of declining NAEP scores, in Utah and nationally, as the receipt (Horvath, 2025). Trent Cash and his colleagues, writing in Trends in Cognitive Sciences, found something very telling: skills handed off to AI, whether solving a math problem or working through a difficult text, tend not to be retained. "If we offload a specific skill to AI, we're probably not going to retain that skill particularly well," as Cash put it.

The Relationship

Neil Selwyn's research keeps this issue from collapsing into either alarm or denial. Selwyn's argument isn't that AI can't replicate functions of teaching. It's that the social, emotional, and relational qualities of a human teacher aren't overhead sitting on top of the real work of instruction. They are the essential aspect of the work (Selwyn, 2019). A tool that quietly erodes the teacher-student relationship has failed, even if it raises a test score. A tool that strengthens that relationship, or frees a teacher's time to invest more in it, is doing what it should.

What the Evidence for Intentional Use Actually Shows

The Cash and Horvath research doesn't settle the question alone, because how AI gets used matters as much as whether it gets used. Gregory Kestin and his colleagues at Harvard ran a randomized controlled trial comparing a purpose-built AI tutor against an active-learning physics classroom. Students working with the AI tutor learned significantly more, in less time, and reported higher engagement (Kestin, Miller, Klales, Milbourne, & Ponti, 2025). The researchers are careful to note this doesn't generalize to every task, particularly ones requiring complex synthesis. But it's real evidence that a tool built with genuine pedagogical intention behaves differently than a tool bolted onto a classroom to chase a trend. UNESCO's 2023 guidance makes the same point at the policy level: AI adoption is a values question before it's a technical one, built on human agency, equity, and transparency as clear, non-negotiable constraints, not aspirational language (Miao & Holmes, 2023).

Where This Leaves You

Duri Long and Brian Magerko define AI literacy as a real, buildable competency: the ability to assess AI critically, use it appropriately, and understand where it fails (Long & Magerko, 2020). That's the work in front of every leader reading this, not choosing a side in the doom-versus-hype debate, but building the judgment, in yourself, your staff, and your students, to ask of every AI tool the same question: does this serve the people in this building, or does it distract from what they're actually here to do. Ask that question honestly, every time, and you'll rarely need anyone else's playbook to know what to do next.

Amodei, D. (2026, September 13). Public statement on AI industry pacing and safety [as reported]. Associated Press and syndicated coverage.

Cash, T. N., Kelly, M. O., Macnamara, B. N., & Risko, E. F. (2026). Is AI making us stupid? Trends in Cognitive Sciences. https://www.cell.com/trends/cognitive-sciences/fulltext/S1364-6613(26)00131-2

Horvath, J. C. (2025). The digital delusion: How classroom technology harms our kids' learning, and how to help them thrive again. Self-published; republished 2026 by Harmony/Convergent.

Kestin, G., Miller, K., Klales, A., Milbourne, T., & Ponti, G. (2025). AI tutoring outperforms in-class active learning: An RCT introducing a novel research-based design in an authentic educational setting. Scientific Reports, 15.

Long, D., & Magerko, B. (2020). What is AI literacy? Competencies and design considerations. Proceedings of the 2020 CHI Conference on Human Factors in Computing Systems, 1-16. https://doi.org/10.1145/3313831.3376727

Miao, F., & Holmes, W. (2023). Guidance for generative AI in education and research. UNESCO. https://doi.org/10.54675/EWZM9535

Selwyn, N. (2019). Should robots replace teachers? AI and the future of education. Polity Press.

Edupreneur's Network: Navigating the 2025 Education Landscape

This week's Edupreneurs Network essay lines up directly with this issue's argument, a genuine topical match. Drawing on the 2024-25 Edtech Top 40 Report, the essay covers districts demanding transparent AI systems, clear data privacy policies, and tools to monitor algorithmic bias, the same values-before-technology case this issue makes, just from the vendor side of the table instead of the leadership side. If this week's argument that AI must serve students and staff resonated, this essay shows what that same standard looks like when you're the one building the tool.

"Future Directions in Educational Thought Leadership"

Chapter 10 is the book's own closing chapter, and its opening section, "Responding to Technological Transformation in Education," is built around the same question this issue closes the series with, made sharper now by a year of real AI safety warnings, cognitive-cost research, and public backlash the chapter couldn't have anticipated. Its "Rethinking AI in Education" section traces four primary classroom applications of AI, intelligent assessment, personalized tutoring, adaptive learning, and predictive analytics, and argues that effective thought leaders respond to these shifts with conceptual frameworks, or are they just technical fixes or an automatic, knee-jerk distrust.

This week: Read "Rethinking AI in Education" and "Integrating Digital Literacy and Critical Thinking" in Chapter 10. Then ask yourself: of the four AI applications named there, which is your system actually ready for, and which are you adopting before you're ready?

If you’re a decision maker at your company, you need to be on the bleeding edge of, well, everything. But before you go signing up for seminars, conferences, lunch ‘n learns, and all that jazz, just know there’s a far better (and simpler) way: Subscribing to The Deep View.

This daily newsletter condenses everything you need to know about the latest and greatest AI developments into a 5-minute read. Squeeze it into your morning coffee break and before you know it, you’ll be an expert too.

Subscribe right here. It’s totally free, wildly informative, and trusted by 600,000+ readers at Google, Meta, Microsoft, and beyond.