iiinova — where your digital self keeps evolving

Milan Toma @tomamil.bsky.social · 05/05/2026

Another week, another “AI outperforms doctors” headline; this time from a Science study published on April 30, 2026. Why these viral AI-beats-doctors stories are dangerous, explained: youtu.be/acP_YelM-i4

youtu.be

Viral AI-Beats-Doctors Studies are Dangerous

Milan Toma @tomamil.bsky.social · 05/05/2026

100

Thank you for mentioning our study. 🥰 Merci d'avoir mentionné notre étude.

Milan Toma @tomamil.bsky.social · 20/04/2026

AI ≠ chatbots. In healthcare, that confusion can be dangerous. Fluent language is not diagnostic competence. My latest video explains what (current and future) healthcare professionals must understand now. 🎥 youtu.be/FT21oU46u9A #MedEd #AIinHealthcare #HealthcareAI

youtu.be

Artificial Intelligence Is Not Just Chatbots: What Healthcare Education Must Teach Now

Milan Toma @tomamil.bsky.social · 17/04/2026

000

Proud to be part of @nyitlibrary.bsky.social's inaugural Faculty Book Launch for National Library Week, featuring 3 faculty and their new books. I’ll be introducing my book, 'Diagnosing AI'. Excited to celebrate scholarship and community with my colleagues! #NationalLibraryWeek #NYIT #DiagnosingAI

041

Milan Toma @tomamil.bsky.social · 13/04/2026

Healthcare AI, medical education, clinical diagnostics, chatbots, task-specific AI, patient safety, machine learning in medicine — I explain why linguistic fluency must not be mistaken for diagnostic competence and why healthcare education is now at a crossroads. www.youtube.com/watch?v=2aRj...

HealthCare Programs at Crossroads: Evaluation of AI in Medicine

Milan Toma @tomamil.bsky.social · 25/03/2026

020

Representing NYITCOM at #AIWeekNY, the festival celebrating innovation across the AI ecosystem in NYC! Discover more: pulse.nyc/ai-week/ My talk is on May 11, where I’ll discuss task-specific AI for clinical diagnostics. Register here: luma.com/e9ml3r51

luma.com

Diagnosing AI · Zoom · Luma

Join Dr. Milan Toma, author of Diagnosing AI: Evaluation of AI in Clinical Practice, for a live session exploring one of the most pressing questions in modern…

000

Milan Toma @tomamil.bsky.social · 18/03/2026

Curious about trustworthy AI in medicine? 🩺🤖 Check out my free YouTube course and learn to spot the pitfalls behind “impressive” AI accuracy, avoid data leakage, and demand real-world validation. Perfect for clinicians, students, and researchers! Watch: www.youtube.com/playlist?lis...

Diagnosing AI: How to Separate Hype from Real-World Reliability in Medical Machine Learning - YouTube

Are you ready to look beyond the headlines and truly understand what makes artificial intelligence trustworthy in medicine? This is a course for clinicians, ...

001

Milan Toma @tomamil.bsky.social · 10/03/2026

I created a podcast episode to explain the accuracy paradox and its connection to medical AI. 🩺🤖 Discover how this concept impacts AI performance in healthcare. Listen here: open.spotify.com/episode/7MmC...

open.spotify.com

The Accuracy Trap

000

Milan Toma @tomamil.bsky.social · 01/12/2025

What can a 2017 colonoscopy study teach us about using AI diagnostics safely in 2025? I explain this critical analogy and provide a full guide for physicians to evaluate AI tools in my latest presentation. Watch the full video: www.youtube.com/watch?v=zmTK...

A Guide for Integrating AI into Clinical Diagnosis

Milan Toma @tomamil.bsky.social · 01/12/2025

000

I’ve created three videos on AI-Assisted Medical Diagnostics, each focusing on a specific area: Technical: www.youtube.com/watch?v=ycy7... Clinical: www.youtube.com/watch?v=zmTK... Practical: www.youtube.com/watch?v=HZ_z...

Training Clinically Actionable Machine Learning Models

Milan Toma @tomamil.bsky.social · 20/10/2025

000

My babies in hard covers 🥰

Two books by Milan Toma; The Economy of Distrust: Medical misinformation and the business of conspiracy; AI-Assisted Medical Diagnostics: A clinical guide to next generation diagnostics.

000

Milan Toma @tomamil.bsky.social · 18/10/2025

My dog doesn’t like the book. He knows it’s the reason I wasn’t playing with him enough.

000

Milan Toma @tomamil.bsky.social · 14/10/2025

From our latest publication, "The objective of this study was not the maximization of predictive performance, but rather the demonstration of the essential role of clinical expertise in the development of reliable ML models for healthcare." DOI: 10.3390/surgeries6040087

000

Milan Toma @tomamil.bsky.social · 18/09/2025

Excited to share my new book: AI-Assisted Medical Diagnostics! It is a practical guide for clinicians; teaching you how to spot trustworthy AI/ML tools vs. hype in medical diagnostics. Learn to evaluate, integrate, and question AI solutions for safer, smarter patient care.

000

Milan Toma @tomamil.bsky.social · 03/06/2025

As demonstrated in our latest NYITCOM publication, we advocate that future ML-driven diagnostic studies adhere to the recent @fda.gov's recommended credibility assessment by analyzing training dynamics alongside traditional performance metrics. DOI: 10.3390/info16060432

020

Milan Toma @tomamil.bsky.social · 01/01/2025

Most ML studies lack external validation, raising concerns about the reliability of their reported performance. Our latest NYITCOM study demonstrates how without external testing, models do not perform as expected in real-world clinical settings. DOI: 10.3390/surgeries6010003

Milan Toma @tomamil.bsky.social · 26/12/2024

Our first article of 2025, "Impact of Convulsive Maternal Seizures on Fetus Dynamics", is out in the International Journal for Numerical Methods in Biomedical Engineering; Volume: 41, Issue: 1. DOI: 10.1002/cnm.3901. Explore our insights into maternal-fetal health at NYITCOM.

Milan Toma @tomamil.bsky.social · 14/12/2024

Doctors' diagnoses were correct 74% of the time without AI, 76% with AI assistance. If relied solely on AI, accuracy soared to 90%, demonstrating “doctors’ unwavering belief in a diagnosis they made, even when a chatbot suggests a better one.” (Not suggesting they should let AI make the diagnoses.)

nytimes.com

A.I. Chatbots Defeated Doctors at Diagnosing Illness

A small study found ChatGPT outdid human physicians when assessing medical case histories, even when those doctors were using a chatbot.

000

Milan Toma @tomamil.bsky.social · 12/12/2024

They decorated everyone’s office door. Finding a workplace with a positive culture is essential for our well-being and productivity. A healthy environment promotes collaboration, creativity, and growth. Lucky me to have found such environment at NYITCOM. 🥰

030

Milan Toma @tomamil.bsky.social · 03/12/2024

Student and professor with their poster on Deep Learning Assessment of Knee Osteoarthritis presented at NYITCOM’s First Annual Biomedical Research, Innovation, and Imaging Center Research Retreat. I’m constantly impressed by our research students’ dedication, creativity, and passion for discovery.

Milan Toma @tomamil.bsky.social · 21/11/2024

Psychosis is a serious mental disorder characterized by a disconnection from reality.

262

Milan Toma @tomamil.bsky.social · 19/11/2024

Representing New York Tech’s College of Osteopathic Medicine at the 8th International Conference on Computational & Mathematical Biomedical Engineering (at George Mason University) with talk titled, “Biomechanical Effects of External Loading on Uterus Dynamics and Fetal Stress”.

020

Milan Toma @tomamil.bsky.social · 18/11/2024

Huge thanks to our New York Tech College of Osteopathic Medicine student Daniel Nasef for sharing our research on AI-driven classification of diabetic patients at the 15th Annual Machine Learning Symposium, hosted by the New York Academy of Sciences at the New York Academy of Medicine.

000

Milan Toma @tomamil.bsky.social · 17/11/2024

Our research paper, "Regressive Machine Learning for Real-Time Monitoring of Bed-Based Patients," has been published in Applied Sciences! It introduces an ML model that enhances real-time monitoring of bedridden patients. A special shout-out to the NYITCOM student co-authors!

Milan Toma @tomamil.bsky.social · 16/11/2024

Representing New York Tech’s College of Osteopathic Medicine at the 18th Long Island Systems, Applications and Technology Conference, with our paper, "Algorithm Selection and Data Utilization in Machine Learning for Medical Image Classification," copublished with student physician Mr. Gazi Hussain.