AI 4 Healthcare Newsletter

Welcome to our newsletter, a meticulously curated platform intended to deliver the most recent news, breakthroughs, and published research from the exciting crossroads of AI and Healthcare.

Welcome to our inaugural edition! We trust you'll find the content enriching. Please share this resource with friends who could benefit from it. Your feedback is invaluable; let us know how we can refine this newsletter to become your most beneficial guide on your AI journey and educational pursuits.

AI News in Healthcare

• The Washintong Post. The Checkup With Dr. Wen: What cautious adoption of AI in medicine looks like.

• The Hill. Who pays the bills when AI kills?. Why this is important: a very important question needs to be answered, if an AI model makes mistakes that can change patient outcomes, who will be responsible?

• TNW. AI in healthcare could exacerbate ethnic and income inequalities, scientists warn.

• Backer’s HealthIT. Google, Microsoft face off in AI healthcare race. Why this is important: there will be a war between all big Tech companies over who owns the healthcare AI. This will bring billions of dollars to the winner.

• Fox News. AI app helps aging adults manage their prescriptions with one photo: ‘Your personal health assistant.

• American Medical Association. ChatGPT, AI in health care and the future of medicine with AMA President Jesse Ehrenfeld, MD, MPH.

• EY. Generative AI's potential to transform the healthcare sector.

• American Council on Science and Health. Thinking Out Loud – Artificial Intelligence Comes For Healthcare.

• SciTeckDaily. Artificial Intelligence Unlocks New Possibilities in Anti-Aging Medicine

AI News in the World

• Bloomberg. Apple Tests ‘Apple GPT,’ Develops Generative AI Tools to Catch OpenAI.

• Meta AI, releases Llama 2, a rival to ChatGPT that is open source and free for commercial and personal use. Why this is important: the model is open source and is available for personal and commercial use. Estimates suggest that it is as good as GPT 3.5.

• The New York Times. OpenAI is worried that some users could use its new visual search capabilities can for facial recognition.

• OpenAI.has introduced a significant update to ChatGPT called 'Custom Instructions' that enables users to have control over responses/expanded the message limit for GPT-4 from 25 to 50 every 3 hours/currently available in Beta for Plus users globally, excluding the UK and EU. Why this is important: these customizations will enhance the usability and flexibility of the chatbot.

• The New York Times. Google is testing an AI tool that can write news articles.

• XAI. Elon Musk launches his new company “xAI” with a a goal to o understand the true nature of the universe. It will be interesting to see what this company will do in the next few months, many think it is a rival to ChatGPT. Why this is important: Elon never puts his mind and energy into something without success, the unanswered question now is: what is understanding the universe means? let us wait and see.

• Fortune. Generative A.I.: 4 things executives should do to future-proof their strategy. Here they are: 1) Focus company resources on using generative A.I. to create competitive advantage, 2)Have a centralized data strategy, 3)Treat the choice of LLM like a choice of strategic partners, 4)Experiment to predict the future of strategic workforce planning.

• Harvard Business Review. 3 Steps to Prepare Your Culture for AI. 1) Choose curiosity over fear, 2) Embrace failure, 3) Become a learn-it-all. You can have the best AI tool in the world, if you culture is not ready to accept it, you will fail.

Impactful Publications

• A study published in journal Nature built a diffusion model (RF diffusion) that can accurately predict the structure a of hundreds of designed symmetric assemblies, metal binding proteins and protein binders. The accuracy of RFdiffusion is confirmed by the cryo-EM structure of a designed binder in complex with Influenza hemagglutinin which is nearly identical to the design model. The impact of this study on drug development can be huge. Make sure to check this paper out.

• Researchers from Stanford and UC Berkeley have published a study suggesting a potential decline in ChatGPT's performance. They observe that certain outputs generated by ChatGPT seem to be deteriorating over time. The figure provided below eloquently encapsulates their findings; as they say, a picture is worth a thousand words. It remains unclear whether this trend reflects an actual decline in performance or reduced usage during the summer season, potentially driven by individuals allocating more time for beachside relaxation. 😄

• A study published in the Lancet Digital Health showed that a deep learning algorithm can predict cardiac functions such as ejection fractions and others from CXR with variable accuracies. Is this important: Yes and no. While it's intriguing to predict these outcomes from X-rays, it does not negate the need for an echocardiogram and other tests for confirmation. The question remains whether the cost of deploying this model in clinical practice will be justified, or if this is merely an intriguing paper, published in a high-impact journal, that will never see real-world application.

• A study published in Nature Human Behavior showed that incorporating the distribution of human expertise by training unsupervised models on simulated inferences that are cognitively accessible to experts dramatically improves (by up to 400%) AI prediction of future discoveries.

• A comprehensive and thoughtful review published in Nature Biomedical Engineering regarding algorithmic fairness in artificial intelligence for medicine and healthcare. Make sure to check it out.

• A study published in JAMA Ophthalmology showed that a Chatbot correctly answered approximately half of the multiple-choice questions in the free OphthoQuestions trial.

• Interesting study published in Nature Machine Intelligence showed that the developments in current computing hardware platforms, storage infrastructure, networking and domain expertise in healthcare cannot keep up with the exponential growth in resources demanded by the AI/ML models.

  Summary of important take away messages from this research:

  1) The article discusses the growing resource sustainability issues in the field of AI and machine learning (AI/ML) for healthcare. These issues include energy consumption, storage, computing power, networking, and domain expertise. 2) The authors note that AI/ML models are experiencing an exponential increase in their size and demand for resources, which is outpacing the developments in current computing hardware platforms, storage infrastructure, networking, and domain expertise. 3) The article highlights the need for algorithm/system innovations to address these sustainability issues and outlines future directions to proactively tackle them. 4) The authors argue that addressing bottlenecks in algorithm and system design with sustainability awareness and promoting collaborations between academia and industry are key to resolving emerging resource sustainability issues. Please take a look at one of interesting images in the study below

Interesting AI tools

Tome, a new AI tool that can generate slides from prompts. I have not tried it yet but it looks interesting (I have no affiliation or received any financial incentives with this company) I just found this product interesting.

Check out our courses from AI 4 Helthcare

ChatGPT for Healthcare. Discover all you need to know about ChatGPT and its applications in healthcare. Join over 2000 students worldwide who have already enrolled in this course.

No code-low code Machine Learning for Healthcare. This is the only course that provides hands-on examples of machine learning in healthcare. Start building real models immediately upon completion of the course.

Artificial Intelligence and Machine Learning in Healthcare (The Basics). Learn everything you need to know about the basics of AI application in healthcare, explained in simple terminology, without requiring any coding or mathematical experience.