🔍 Spotlight #10: 💊 Antibiotic Discovery, 🤖 Penny Assisting Cancer Care, 🩺 Philips IGT, ⚠️ AI Risk
AI Health Hub, 07/06/2023
💊 New superbug-killing antibiotic discovered using AI
News
Great article by BBC, especially if you are not too familiar with this research field as myself. This piece explains very simply the research findings of a recently published paper in Nature Chemical Biology (open access) on machine learning for antibiotic discovery. The group targeted the most problematic species of bacteria, Acinetobacter baumannii, which can infect wounds and cause pneumonia. After training the AI, they used the system on 6,680 compounds, and in an hour and a half, it produced a shortlist. Laboratory testing showed 9 potential antibiotics. The potential of AI in this field is to significantly accelerate the search for novel antibiotics, decrease costs, and reduce side effects and drug resistance as the experimental antibiotic was very precise and had no effect on other species of bacteria.
🤖 Penn Medicine uses AI chatbot 'Penny' to improve cancer care
News
The development of oral anti-cancer drugs, including cytotoxic chemotherapies that patients can self-administer at home comes with some concerns about compromised efficacy if patients take more or less than the prescribed doses. A research group at the University of Pennsylvania Health System is carrying out a study with the aim of improving medication adherence, monitoring and managing toxicities related to oral chemotherapy for patients with gastrointestinal cancer by means of an AI chatbot. After some searching, I was able to find the link to the clinical trial if you want to check it out here. The news article mentions some results which I could not find published anywhere, but the group claims that over 500 symptom-related text messages were exchanged and approximately 98% were correctly interpreted by Penny, the chatbot. Nearly 4,000 medication-related-text messages were exchanged and approximately 93% were accurately interpreted by Penny. The clinical trial is still ongoing and no publications are available yet.
🩺 Innovation in Image-Guided Therapy with Dr. Atul Gupta
Podcast
Tune into this very interesting 42-minute episode of the Less Invasive podcast with guest Dr. Atul Gupta, time will fly by! Dr. Gupta is an interventional radiologist and Chief Medical Officer at Philips Image Guided Therapy (IGT). I find it always interesting to hear about the background and journey of medical doctors joining the industry. Besides sharing more about his medical career and current role, Dr. Gupta touches upon the current innovations carried out by Philips (IGT) with regard to Augmented Reality, AI, and robotics. His opinion on AI is that it should not be “artificial intelligence” but “assisting intelligence”, to make better use of the amounts of data. Today is being used to reduce the time of intervention for stroke patients with early warning systems, speeding up imaging, and for cardiac disease by embedding AI in patient monitoring systems, overall helping us intervene before anything happens. He also mentions the need for caution when training the models as we should take into account patient differences. Similarly to previously shared opinions, he believes there is a need for more collaboration between regulators, payers, government, and industry. Data representativity and bias are to be considered. Finally, he speaks about studies on health economics to be carried out in parallel with the clinical evidence that can prove cost savings at the same time as better outcomes for the patients, as healthcare funds shrink around the world.
⚠️ Statement on AI Risk
Statement
I am linking here the official statement from the Center of AI Safety but also their LinkedIn announcement as it includes links to extensive blog posts by three of the signatories Sam Altman, Yoshua Bengio, and David Krueger (Cambridge) on their concerns. The statement is quite brief as they mention the signatories have different views on AI. It reads:
Mitigating the risk of extinction from AI should be a global priority alongside other societal-scale risks such as pandemics and nuclear war.
You can explore the list of signatories, and there is also a very nice feature that allows you to filter between “AI Scientists” and “Other Notable Figures“.
🤩 Another week, another 🔍 Weekly Discovery!
As I mentioned in one of the last editions, I started a new job last week (🥳). A lot of changes, reading materials, skills to learn, and meeting new people meant less time to dedicate to this newsletter. As much as I enjoy preparing the weekly and being on time, I decided to be kind to myself and be more flexible with the release of this newsletter. I will still keep the weekly cadence since I do continue reading and listening to podcasts, and I have plenty of commuting time for now 😅. The day of the week will be flexible and, hopefully, I will find a stable routine soon!
❗️If you recently subscribed, take time to explore the last editions too, you might find podcasts, scientific articles, or recent news that could interest you.
🤗 I appreciate all of you readers, and you can always leave comments for suggestions, notes of improvements, or if you found it useful!