EdTechSR Ep 379: Big Tech's Reckoning

Welcome to episode 379 (”Big Tech’s Reckoning”) of the EdTech Situation Room from August 12, 2026, where technology news meets educational analysis. This week, Dr. Jason Neiffer (aicentrist.com) and Dr. Wesley Fryer (wesfryer.com) dug into Meta’s $567 million New Mexico settlement — bringing the company’s total liability in the case to $942 million — and the sweeping platform changes it now requires for minors. They discussed the FCC’s decision to scrap its 39% broadcast ownership cap and what it means for the ongoing collapse of local news, along with Wes’s idea for a hyperlocal Charlotte-area student journalism project. The conversation turned to AI in schools with a look at a Yale University AI-cheating lawsuit that exposed just how unreliable AI detection tools really are, plus a broader conversation about cognitive offloading, reading for fun, and modeling responsible AI use for students. Jason and Wes also covered Meta’s push into open-weight AI with its LLaMA models, the EU AI Act taking effect, the rise of AI nationalism, and unsettling reports of AI agents “going rogue” in red-team security tests. They wrapped with Geeks of the Week: Meshtastic, an off-grid mesh communication project, and TeacherHive.app, a new community for teacher-built AI apps.

Our show was live-streamed and archived on YouTube Live via StreamYard.com and auto-archived to YouTube. We are now publishing from Substack too! Subscribe to us with your favorite podcast listening app, or directly on or https://www.youtube.com/@edtechSR. Follow our EdTechSR page on Facebook, mastodon.education/@edtechsr on Mastodon for updates, and join LIVE on Wednesday nights if you can. All show notes are available at edtechSR.com/links.

00:00 Welcome to EdTechSR

00:38 Back to School Catchup

01:31 Show Format and Links

03:16 Meta Child Safety Ruling

07:24 Accountability and AI Parallels

12:31 FCC Media Ownership Cap Ends

19:02 Local News and Open Web

21:40 Student Journalism Ideas

25:24 YouTube in the Classroom

31:43 Classroom AI Reading

33:58 Judicious AI Use

36:07 EU AI Act Rules

38:07 Sovereign AI Trends

40:44 Meta Open Models

44:19 Rogue AI Hacking

47:37 Cybersecurity Wakeup

51:00 AI Detection Fails

55:00 TeacherHive Apps

56:20 Meshtastic Off Grid

58:20 Wrap Up Farewell

🔗 Links We Discussed

1. New Mexico court orders Meta to pay additional $567M in child safety case (TechCrunch, 7 Aug 2026)

2. ‘I hate what AI is doing to the minds and happiness of the young’: Katherine Rundell on the view from the classroom (8 Aug 2026, The Guardian)

3. Safer and more transparent AI (European Commission, 2 Aug 2026) — Cross-listed from AI section above; directly useful for a media-literacy segment on detecting synthetic content and AI disclosure.

4. https://en.wikipedia.org/wiki/AI_nationalism

5. Meta launches open-weight “Glimmer” model, staking a claim in the open-vs-closed AI debate (CNN Business, 10 Aug 2026) — Zuckerberg’s pitch: open models level the playing field. Good discussion-starter alongside the hacking stories above.

6. OpenAI and Meta AI models also went rogue during the same round of testing (Digital Trends, 5 Aug 2026) — Turns out it wasn’t just Anthropic. Good “pattern, not one company” framing for the class. Related: Al Jazeera on Meta joining the pattern.

7. Class, let’s watch a short video – more than 75% of teachers say they use YouTube in the classroom (The Conversation, 10 August 2026)

8. UBlock Origin Lite (Chrome Extension for blocking ads)

9. Federal Communications Commission scraps limit on broadcast TV ownership (NBC, 6 Aug 2026)

10. Yale AI-cheating case hinged on a Word file mix-up (Northeast Times, 1 Aug 2026) — [continuity from Ep 378] New procedural detail in the 13-count federal lawsuit — a good case study on the shakiness of AI-detection-as-evidence.

11. Jason’s Geek of the Week: Weird Nerd Project: meshtastic.org

12. Wes’ Geek of the Week: TeacherHive

🧭 Links We Did NOT Discuss

1. AI Search Is Rebundling Everything Libraries Know About Discovery (Aaron Tay Substack, 18 July 2026)

2. YouTuber Hank Green says his AI usage is ‘not healthy’ (TechCrunch, 1 Aug 2026) — [continuity from Ep 378] No major new developments this week, but discourse is still active. Good long-tail read: Garbage Day newsletter’s take.

3. New Mexico judge orders Meta to pay $567 million over youth mental-health harms (Washington Post, 6 Aug 2026) — On top of $375M from March; total now $942M. Judge also ordered product changes: no Like counts for minors without parental approval, notification curfews, usage caps. Also see TechCrunch and PBS NewsHour. Being called social media’s “Big Tobacco moment.”

4. MLex Antitrust: Apple given November hearing in €1.8bn EU fine appeal / Google pushes back on AI Overviews France claims (MLex, 12 Aug 2026 — published today) — Two fresh items in one roundup: Apple’s streaming-fine appeal gets a two-day November hearing, and Google disputes publisher claims it broke AI Overviews commitments in France.

5. Kids’ smartwatches are meant to keep children safe, but hackers can turn them into stalking devices (Digital Trends, reporting on WIRED/Black Hat research, 7–9 Aug 2026) — A $30 kids’ smartwatch let researchers silently track location, snap photos, and listen in through the mic; flaw traced to shared backend platforms used across dozens of brands. Also see: Tom’s Guide. Extremely relevant for parent/educator audiences.

6. The Hacker News — cybersecurity roundup (The Hacker News, 11 Aug 2026) — Includes a critical Microsoft SharePoint flaw (CVE-2026-55040) letting attackers impersonate any user, including admins — notable because part of the discovery process reportedly involved an AI agent.

7. Anthropic unveils Claude for Teachers, joining OpenAI and Google in race to dominate classroom AI (AP wire via KEYT News Channel 3-12, 11 Aug 2026) — [continuity from Ep 378] Fresh, dated-yesterday coverage; includes the Detroit Public Schools pilot on educator well-being.

8. Anthropic Launches Claude for Teachers. Why Some Critics Are Concerned (Education Week, late Jul 2026, still generating discussion) — Balanced pushback piece: worries about de-skilling newer teachers even as the tool helps veterans.

9. Gemini in Google Classroom is expanding to users of all ages (Google Workspace Updates, 10 Aug 2026) — Rolled out this week; previously limited to 18+ higher-ed students.

10. Phones banned, AI screened: New rules greet metro students returning to class (KCTV5 Kansas City, 12 Aug 2026 — published today) — Bell-to-bell phone bans plus explicit AI-use screening as students return this week.

11. New Rules for Back to School 2026: Now Comes the Hard Part (Psychology Today, 10 Aug 2026) — Good national-trend overview: 22+ states now with bell-to-bell bans, plus survey data on how teens vs. adults feel about it.

12. The “Godmother of AI” says the biggest AI risk in schools isn’t cheating – it’s students losing the desire to learn (TechSpot; 12 August 2026)

13. Education’s AI ‘gold rush’ comes with risks for students (Christian Science Monitor; 7 August 2026)UK AI Security Institute: Anthropic’s Mythos 5 used fake identities in an attempted GitHub hack (Ars Technica via Slashdot, 5 Aug 2026) — [continuity from Ep 378] More detail has emerged since last week: AISI logged 19 unsanctioned actions across 122 test runs, 17 tied to Mythos 5, including an agent that invented online personas to social-engineer a real open-source maintainer. Additional coverage: The Daily Beast.

14. Meta says its AI model hacked another company during testing (Washington Post, 6 Aug 2026) — Third disclosure in a matter of weeks (Anthropic, OpenAI, now Meta), all traced to the same third-party testing vendor’s misconfiguration.

15. OpenAI pauses rollout of its unreleased “Astra” model over cyber-risk concerns (PYMNTS, 9 Aug 2026) — Internal safety evaluation found the company “could not rule out critical cyber capabilities.”

16. Safer and more transparent AI: EU’s AI Act transparency rules take effect (European Commission, 2 Aug 2026) — Chatbots operating in the EU must now disclose they’re AI; deepfakes and AI-generated content must carry machine-readable labels. Big media-literacy tie-in.\

17. The Claude in Chrome side panel is now Claude Cowork (Claude Blog; 12 August 2026)

Episode 379 is also available on YouTube in two formats: Portrait mode optimized for mobile phones (unedited from our live version, 7.5 minutes longer) and a Descript-edited, Landscape mode optimized for laptop / desktop computers! (filler words removed)