October 18, 2017
GET READY FOR ARTIFICIAL INTELLIGENCE. IT’S COMING – AND NOW!
Artificial intelligence: friend or foe? Are we facing our own impending doom by this proliferating technology we’ve been warned about from countless apocalyptic sci-fi movies, or should we welcome this humble and helpful servant to faithfully carry out the menial tasks that we can’t be bothered to do anymore? Anthony Panetta and Brett Miles of Pfizer helped us gain some clarity on these important questions at October’s PMCQ event.
Brett opened the meeting by covering a few points of business:
- The social media sites for the PMCQ are up and running. Please follow us!
- The PMCQ has a new educational sponsor: Lundbeck.
- There are also new corporate sponsors: Ashfield Healthcare and EMREACH Inc.
- The next event will be on November 22nd and will be a great presentation about preparing pharma professionals for the pharma industry of the future.
- “The Art of Leadership” is another upcoming event and the PMCQ offers a promotional code if this is an event you are interested in attending.
- A survey will be going around asking for feedback about an upcoming event. Please take the time to let us know what you think!
- Thank you to Fusion MD for doing the creative for this event and Pfizer for their sponsorship of the event.
Anthony Panetta then introduced the night’s speakers, Dr. John Reeves, MD (Chief Medical Officer) and Lexi Kaplin (Chief Product Officer) of emojiHEALTH. In their business, they use digital technology to transform healthcare. We learned a lot from them!
What we learned
1) How to talk like techies
Here are the key definitions you need to know to sound like a digital nerd:
Artificial Intelligence (AI): an umbrella term for when a computer can mimic human cognitive functions. AI can be visualised as an “ecosystem” in which “Machine Learning” is a sub-category of AI and “Deep Learning” is a sub-category of “Machine Learning.”
Machine Learning: the ability of the computer to take data points and improve upon its basic functions. It’s the computer’s ability to learn by absorbing information.
Deep Learning: the technology that allows a computer to think like a human by mimicking the neural networks of a human brain and having rational thought processes.
2) AI is already mainstream
Computers are already able to do tasks like:
- Understand written text, even in colloquial form, such as “yep” instead of “yes”. This is called natural language processing.
- Example: A written query to a computer that asks, “What are the side effects of drug X?”
- Interpret verbal communication, which involves converting a spoken query into text, searching for an answer to the query, and then converting the text answer to speech in order to respond to the query. This is called speech recognition.
- Example: Asking the computer verbally, “What are the side effects of drug X?”
- Recognize images and even text within images. This is called image recognition.
- Example: Uploading an image of a pill bottle and having the computer read the label.
Researchers are working on improvements in machine learning, but it is also already readily available. Deep learning is still in development, although there are a few applications of it.
3) We need AI because the human attention span is shorter than that of the goldfish
AI is a practical solution for consumers because it makes engagement scalable and facilitates repetitive tasks. As humans, we don’t like to wait to find someone to give us the information we seek, and we don’t have the time to do boring tasks that something else can do. According to Lexi, by 2020, 85% of customer interactions will be managed without a human.
Pharma and healthcare will be affected in 4 key ways:
- Repetitive tasks
- Example: An app that allows patients ttake images of skin conditions and get a real-time diagnosis
- Example: Cognitive assistants for physicians to analyze radiology images to spot simple medical defects faster and more reliably, allowing the physician to focus on more complex cases
- Example: An assistant that uses voice recognition to answer a physician’s request to look up medical information, record clinical notes, verify dosing information, check schedules, etc.
- Precision medicine
- Tapping into the power of predictive analytics to analyze medical interactions at a molecular level. Applications include genetic interactions, vaccine creation, visual biopsies, etc.
- Example: Analysis of genotypes and phenotypes to understand diseases before they occur
- Example: Drug discovery
- Example: Identification of biomarkers to perform and classify a visual biopsy in real time at a scale that humans cannot see
- Example: Drug compatibility
- Example: Assisted surgery
- Data mining
- Example: The use of predictive analytics to understand huge data sets to achieve the best possible treatment for patients in time
- Example: An app that checks a user’s symptoms against a database using speech recognition, and then offers an appropriate course of action
- Example: Analysis of invoices to determine if a doctor, clinic or hospital repeatedly makes mistakes in treating any condition, in order to help clients improve their service and avoid unnecessary hospitalizations of patients
- Customer engagement
- Example: Chat bots that communicate with customers and give them the information that they want in a bite-sized, digestible, visual format, and give it to them on demand and friction-free
4) Pharma can develop chat bots now
Here are some ideas...
- Patient bots to educate patients about a condition or product
- Clinical trial bots to recruit and onboard participants
- HCP bots to help sell products
- Internal bots to communicate internally to your sales force
5) AI will NOT steal your job
AI needs humans and humans need AI. The magic combination is when AI does 85% of the work (the time-consuming and repetitive tasks), allowing humans to focus and master the remaining 15%.
Freelance Medical Writer
Cell: (514) 605-5109