DOWNLOAD THE PDF BULLETIN HERE

Ted Jedynak, the AI for Allied Health Guy · Helping Clinic Owners Implement AI Safely & Effectively for Better Clinical Outcomes.

WTF Happened in AI Allied Health This Week? | Ted's Friday Five — 3 July 2026

TEDucation | EP. 005

Harvard Medical School just launched an eight-week AI in healthcare program. USD $3,000. That's over five thousand, two hundred Australian dollars.

I looked at the curriculum. There's a lot of overlap with what I teach. The difference is who it's built for — and this week, that difference matters.

Here are the five things in AI and allied health that actually moved the needle this week.

1. Stanford and Harvard Just Proved the AI+HI=CI Thesis

The ARISE network — a joint research collaboration between Stanford and Harvard — released the first rigorous, real-world audit of how clinical AI actually performs. Not in a controlled benchmark. In practice.

The headline stat: the best-performing AI models make 12–15 severe clinical errors per 100 cases. The worst? Over 40.

And the failure mode isn't happening at the exotic fringes of clinical work. AI breaks down when it faces uncertainty, incomplete information, or multi-step decisions — which describes the majority of what allied health practitioners deal with in every ordinary consultation.

The report also confirms something else: AI adoption in healthcare is being driven primarily by workforce burnout and documentation load. Practitioners aren't reaching for AI because they want to be automated out of a job. They're reaching for it because they're drowning.

What this means for you: My entire Clinical Intelligence Program is built on the idea that AI plus Human Intelligence equals Clinical Intelligence. The human in the loop isn't optional — it's what makes AI safe in a clinical setting. Stanford and Harvard just peer-reviewed that argument.

Save that stat. 12–15 severe errors per 100 cases. You'll need it the next time someone tells you AI will replace clinicians. It won't. But it will replace the clinicians who refuse to learn how to work with it properly.

2. The UN Started Formal Global AI Governance Talks This Week — Australia Has No Dedicated AI Law

The United Nations kicked off formal Global Dialogue on AI Governance on 6 July 2026 in Geneva. The World Health Organisation is co-chairing the health-specific workstreams. Member states are now formally negotiating international frameworks for managing AI risk in clinical settings.

Meanwhile in Australia: we have no dedicated AI law. None. Zero.

What we have is a patchwork — TGA for devices, AHPRA for professional obligations, Privacy Act for data. It's principles-based. It's increasingly thin compared to what other jurisdictions are building.

Here's what every major international framework converging right now — EU AI Act, UK AI Safety frameworks, US executive orders, now UN-level negotiations — has in common: they all mandate mandatory human oversight of AI in clinical settings.

Which is, again, AI+HI=CI.

Every country writing AI law is essentially legislating the concept that the human has to stay in the loop. That's a regulatory tailwind for what the CI Program teaches. And it's a signal to Australian clinic owners that the "she'll be right" approach to AI in your practice has a use-by date.

Get ahead of it. Don't wait for the law to force you.

3. Harvard Charges $5,200 for This. Here's What That Actually Buys You.

Harvard Medical School runs an eight-week AI in healthcare program through Executive Education. Online. Live faculty sessions. USD $3,000 plus — over $5,200 Australian.

It's excellent. Genuinely. Built for hospital C-suite executives, health system leaders, and innovators designing enterprise AI strategy.

It will not tell you how to set up Heidi Scribe before Tuesday's patients.

It will not walk you through Australia's Privacy Act obligations for a small allied health practice.

It will not show a solo podiatrist in regional Victoria how to build a clinic that runs without them at the centre of every decision.

The Clinical Intelligence Program is eight weeks. Australian-built. Allied health-specific. Built for the clinic owner who has patients at 8am — not board presentations at 9.

Harvard's capstone involves pitching an AI-first healthcare strategy to a faculty panel. The CI Program's measure of success is simpler: does your clinic run better by Week 8?

If Harvard is the credential you hang on the wall — the CI Program is the thing that changes what happens inside your clinic.

→ tedjedynak.com/clinical-intelligence

4. OpenAI and Microsoft Are Building Dedicated Clinical AI Products — They're Coming to Australian Clinics

Two announcements worth tracking now, before they land at your door.

OpenAI: In January 2026, OpenAI launched ChatGPT for Healthcare — a GPT-5-powered workspace built specifically for clinicians and administrators. Doctor-led testing. Encrypted patient data integration. Live in the US.

Microsoft + Mayo Clinic: In June 2026, Microsoft announced a partnership to co-develop a clinical AI model with Mayo Clinic. Mayo owns the model outright. The plan: license it to health institutions globally.

Australia is not as far off the radar as we'd like to think. We invented Wi-Fi and the cochlear implant. They know we're here.

These tools will land in your inbox as vendor emails, LinkedIn ads, and conference exhibitors within 12 to 24 months. Your patients will start asking about them too.

When that happens, you want to be the clinician who already knows how to evaluate a clinical AI product — is it TGA-regulated? Does it require patient consent? What are your disclosure obligations? — not the one Googling the answers at 11pm.

That's not a small difference. That's the difference between being ahead of your profession and being dragged along by it.

5. Where Do I Start With AI? — New Course, 28 July 2026

If those first four items left you thinking "Ted, that's a lot, I don't even know where to start" — I've heard you.

I've built a course specifically for that moment. It's called "Where Do I Start with AI?" and it kicks off on 28 July 2026. Designed for allied health practitioners and clinic owners who know AI matters but feel overwhelmed by where to actually begin.

No jargon. No theoretical frameworks. I'm a podiatrist. I spent decades explaining to people why their feet hurt. Translating complicated things into language that makes sense is basically the same skill set.

This is not the Clinical Intelligence Program — that's my flagship deep-dive. This is the on-ramp. The thing you do before that. If you've been waiting for a clear starting point, this is it.

28 July 2026.

→ tedjedynak.com/where-do-i-start

The Week in One Paragraph

Stanford and Harvard proved AI needs human intelligence to be safe. UN governance talks signal where the world is heading — and Australia needs to catch up. Harvard's $5,200 AI program is excellent, but it's not built for your clinic. Big tech clinical AI tools are coming to Australia whether you're ready or not. And if you want to get ready without the overwhelm — the door opens 28 July.

Until next time: serve with spirit, consult with care and ramp up your AI today!

— Ted

Ted Jedynak is a podiatrist, educator, and business coach based in Adelaide, South Australia. He is the founder of TEDucation and creator of the Clinical Intelligence Program (AI+HI=CI).

tedjedynak.com | Clinical Intelligence Program