GenHax

This fortnight, read that twice: the kids who built AI designed the anti-cheating tool. Plus: AI makes critical thinking harder. - Two stories that **looked** like they disagreed.

On June 9, the White House honored the winners of the first Presidential AI Challenge. The elementary co-winners from Alcoa, Tennessee built "Homework Helper," an AI tool that guides students through problems instead of handing them answers. A team of 4th and 5th graders built the thing every worried parent wishes existed.

That same two weeks, 54% of K-12 teachers told NPR/Ipsos that AI makes critical thinking harder, and the U.S. Surgeon General warned about AI chatbots and youth well-being. EdWeek's headline put it plainly: "White House Honors AI Challenge Winners as Tech Backlash Grows."

Here's the thing. These stories are not in conflict. The worry is about kids *consuming* AI. The celebration is about kids *building* it. Two different verbs, and only one of them was on stage at the White House.

GenHax runs the year-round version of that verb, for college students and the workforce.

Program managers: NSF TechAccess: AI-Ready America ($224M, up to 56 hubs) closes its Round 1 letter of intent Tuesday, June 16.

Which stat from this fortnight would you put on the fridge? Tag a teacher who's already on the building side.

Big news for every New York family this month.

The SUNY Board of Trustees just made AI literacy a graduation requirement for all 64 campuses. Every incoming undergraduate starting Fall 2026 will complete an AI literacy component as part of general education. Chancellor John B. King Jr. signed off on it April 30. Inside Higher Ed broke the story on May 4.

Boston Public Schools is doing the same in K-12 starting September. Google launched free AI training for 6 million U.S. educators last week. U.S. bachelor's degrees in AI grew 114% in just one year.

The AI literacy floor is being set everywhere. That is good news.

Here is the part we want to add to the conversation.

Knowing what an AI agent is, and being able to ship one, are not the same skill. Literacy is the syllabus. Building is the portfolio.

GenHax is New York based. While SUNY makes literacy mandatory, our learners have already been at the next floor: building real things on NSF-funded knowledge graphs, cybersecurity infrastructure, climate models, and homelessness support tools.

A few that are open right now:
👉 Code Your First AI App or Agent — the on-ramp with mentor hand-holding
👉 US NSF — Proto-OKN knowledge graphs, 90% of past learning engineers improved on their next AI project
👉 SECURE CHAIN, DREAM-KG, Climate Forecast — NSF-grade research challenges

Two deadlines for educators and program managers this week:
📌 Thursday May 21 — aiEDU Community Catalyst Letters of Intent ($25K and $50K grants)
📌 Wednesday May 27 — NSF FINDERS FOUNDRY Planning proposals (up to $50K, requires K-12 students throughout development)

If you are building a proposal for either: we have run the student-in-the-loop co-creation model 25+ times. We can plug in.

Question for the community: SUNY is setting AI literacy as the floor. What is the next floor for your students?

🔗 Live challenges: app.genhax.com/challenges
🔗 SUNY story (Inside Higher Ed): https://www.insidehighered.com/news/student-success/academic-life/2026/05/04/suny-sets-systemwide-ai-policy

Three AI + education developments worth your attention.

1. A new NSF-funded study from MIT finds that project-based learning combined with AI feedback loops increases knowledge retention by 34% over traditional instruction. The research validates what experiential learning advocates have argued for years.

2. Google's newly released AI Overviews for Workspace now include citations and source attribution. For L&D teams, this is a critical step toward building defensible, audit-able training content.

3. OpenAI's o1 model shows significant gains in reasoning tasks, which matters for educators designing authentic problem-solving curricula. o1 vs Sonnet - we'd love your comments. The question: how do we teach students to think *with* these tools rather than around them?

What's the biggest shift you're seeing in how your teams approach AI literacy training?

The Knowledge Graph Conference 2026 starts on Monday at Cornell Tech : 72 hours. Three days. Three moves for you!

**Friday (today): Pick a graph.**

Browse the Proto-OKN challenges at app.genhax.com.

18 NSF-funded knowledge graphs. Pick the one that bothers you most!

DREAM-KG (homelessness, Temple). SECURE CHAIN (software supply chains, Purdue). ClimatePub4KG (climate literature, Temple). SPOKE (drugs + space biology, UCSF + NASA GeneLab). FRINK (federation layer, RPI). And more!

👉 app.genhax.com/challenges

**Saturday: Build a thing.**

Open to non-tech. Block one focused hour. No tutorials. Explore the resources; find out how to explore the proto-OKNs from various KG explorers. Find Gaps. Design ways to bridge them.

**Sunday: Ship a draft.**

Post something visible, as a "Solution". A query that returned the right answer. A graph built from a CSV. A 60-second screen recording. Ship something this weekend, and we'll share your results with The Knowledge Graph Conference as they open on Monday!

Otherwise this week, Boston Public Schools committed to AI fluency for every high school graduate ($1M seed from Paul English). Penn State launched a 1-hour AI Essentials module for 88K students. The OECD reminded everyone that outsourcing to AI without learning leaves no skill behind once the AI is gone.

Literacy is the floor. Building is the ceiling. KGC 2026 is where the ceiling gets shattered.

👉 knowledgegraph.tech

Sources:

• KGC 2026 (May 4–8, Cornell Tech): https://www.knowledgegraph.tech/

• NSF Proto-OKN: https://www.proto-okn.net/

• Boston Public Schools AI fluency: https://www.wbur.org/news/2026/03/26/boston-public-schools-ai-literacy

• Penn State AI Essentials: https://www.psu.edu/news/academics/story/penn-state-launches-new-online-ai-essentials-literacy-course-employees

• OECD Digital Education Outlook 2026: https://www.oecd.org/en/publications/oecd-digital-education-outlook-2026_062a7394-en.html

AI Agent Builders | LLMs • RAG • Knowledge Graphs
Open Source, Local LLM & AI Agent | Ollama LM Studio Hugging Face WebUI MCP
LLMs and Agentic AI
AI Automation & Integration | NLP Solutions | Agentic AI Development | RAG

Three education tech stories shaking things up TODAY:

1. OpenAI released GPT-4.5 fine-tuning APIs designed for educators and institutions. This means universities and schools can now train their own AI models using their own curricula and data without uploading everything to the cloud. Privacy + customization in one move.

2. The National Center for Education Statistics (NCES) just published findings showing that 68% of K-12 schools are now using AI for personalized learning pathways. That's a massive jump from last year. The schools not moving on this are starting to feel the pressure from families asking tough questions.

3. Stanford released research on what they call "vibe coding" and how the style and tone of prompts actually shape the behavior of AI agents. If you're training teachers or students on AI tools, this research is worth your time.

The common thread connecting all three? Institutions that invest in hands-on experimentation, real-world challenges, and taking control of their own AI tools are winning. The passive observers are losing ground.

What's your school or organization doing to stay ahead on this? Share in the comments. We'd love to hear what's working in your classroom or district.