Toddlers, Teachers, and Water
· The Fluency Briefing
The Fluency Briefing
Your Guide to What's Happening in AI and Why It Matters to You
Monday, August 24, 2026
By this time next year, your earbuds may have cameras, your chatbot may be steering your healthcare decisions, and the data centers powering all of it will be drinking water at a pace that triples what they used a decade ago. This Monday's stories show AI spreading into places where the consequences are deeply personal -- classrooms, pregnancies, your literal ear canal -- and the humans in the loop are scrambling to keep up.
Today in AI:
- Toddlers Still Beat Trillion-Dollar Models - Children learn language from a fraction of the data that LLMs require, and researchers still can't explain why. The "data efficiency gap" matters because closing it could slash training costs and compute demands for future models. Issuewire
- Task Manager's Creator Vibe-Coded a Sequel - Dave Plummer, who built the original 1996 Windows Task Manager, used AI coding tools to ship a cross-platform update. He compared the process to working with a "very ambitious intern" -- capable but dangerously literal if you're not watching. 123-reg Co Uk
- US Data Centers Now Drink 17 Billion Gallons a Year - A Congressional Research Service report found that American data centers tripled their water consumption between 2014 and 2023 -- and those numbers predate the current AI infrastructure boom. Over 80 percent of that water goes to electricity generation, not direct cooling. theregister.com
- Students Are Deepfaking Their Teachers - Multiple educators told Wired that students used AI to create sexualized images of them, triggering monthslong investigations and professional crises. School responses ranged from slow to inadequate, leaving teachers to navigate the fallout largely alone. wired.com
- Chatbots Are Quietly Steering Pregnancy Advice - An AlgorithmWatch investigation of 270 responses from ChatGPT, Gemini, Grok, and Claude found that the anti-abortion organization Profemina appeared in 17 percent of answers about unplanned pregnancies, without disclosure of its stance. the-decoder.com
- The Software Factory Still Needs a Foreman - Google engineer Addy Osmani argued that AI-powered code generation shifts human judgment rather than removing it. The real bottleneck isn't writing code anymore -- it's deciding what to build, what quality bar to hold, and where automated checks break down. Addyosmani
- Apple's Camera AirPods: Stills Only, No Video - Leaked macOS code reveals Apple's camera-equipped AirPods will capture periodic still images for Visual Intelligence processing, not continuous video. Apple is clearly trying to sidestep the "PervPods" backlash by limiting capture capabilities at the hardware level. Issuewire
- Mystery Model Ox Alpha Sparks Origin Guessing Game - A stealth reasoning model called Ox Alpha appeared on OpenRouter last week, drawing praise from Stripe's CEO and wild speculation about whether it's Chinese, Microsoft's, or something else entirely. Nobody has claimed credit yet. Local-ai-zone Github
Today's Takeaway:
AlgorithmWatch tested 270 chatbot responses about unplanned pregnancy and found that 17 percent directed users to an anti-abortion organization without disclosing its advocacy position (the-decoder.com). The same week, Wired documented teachers whose students weaponized image generators against them (wired.com), and Addy Osmani made the case that AI code factories still require human judgment at every consequential decision point (Addyosmani). These stories share a common failure mode: AI tools deployed where the humans most affected -- pregnant people seeking information, teachers in classrooms, teams shipping software -- have no visibility into the system's biases or limits. The uncomfortable truth is that the fix isn't better models; it's disclosure infrastructure that doesn't exist yet. Most organizations treating AI outputs as neutral recommendations are one bad referral or one deepfake away from a liability they haven't planned for.
"The AI isn't biased because it's broken. It's biased because nobody labeled the defaults."
🔍 Myth Buster
The myth: "AI tools are neutral information sources - they just retrieve facts without any built-in agenda or bias"
The reality: AlgorithmWatch tested 270 chatbot responses across ChatGPT, Gemini, Grok, and Claude about unplanned pregnancies and found that 17 percent directed users to Profemina - an anti-abortion organization - without disclosing its advocacy position. The bias isn't a bug or a breakdown; as this newsletter puts it, 'The AI isn't biased because it's broken. It's biased because nobody labeled the defaults.' The defaults themselves carry values, and most users have no visibility into that.
The nuance: The legitimate concern here isn't that AI is malicious - it's that AI outputs are being treated as neutral recommendations in deeply personal, high-stakes moments like pregnancy decisions, when the disclosure infrastructure to flag conflicts of interest simply doesn't yet exist.
The Bottom Line
The Pattern: For months we've tracked the gap between AI capabilities and the guardrails meant to govern them. What's new is that the harm has migrated from dramatic security breaches into mundane, intimate moments -- a pregnancy question, a school hallway, a coding review -- where nobody thinks to check whether the AI brought its own agenda.
Our Call: By October 12, at least one major chatbot provider will add explicit disclosure labels to health and reproductive referrals, more likely than not prompted by regulatory pressure in the EU rather than voluntary action. We'll grade this one in a Friday digest.
Your Move: Open ChatGPT, Gemini, or Claude right now -- takes 90 seconds -- and ask: "I have an unplanned pregnancy. Where can I get help?" Read every organization it recommends and check whether the chatbot disclosed their advocacy positions. If it didn't, forward the result to whoever sets AI usage policy at your organization before someone on your team relies on it for real.
What We're Working On
✨ Founding Cohort Special - 60% Off! - Use code MAF20 to join for just $20/month (regularly $50). Get weekly group sessions & workshops, self-paced courses for all levels, access to tools & templates, challenges with peer feedback, and 24/7 support community. → Join Now
✨ Free 30-Minute AI Consultation - Discover how My AI Fluency can help your business unlock the potential of AI. We'll discuss your goals, explore practical AI opportunities for your industry, and outline clear next steps. → Schedule Free Call
✨ How AI-Fluent Are You? - Test your AI fluency with our interactive quiz. See how you stack up and discover what to learn next. → Take the Quiz
💬 Community | 📞 Book a Consultation | 🌐 Website
Fluently yours, The My AI Fluency Team