Pro Yearly is on sale from $80 to $50! »

How can investors spot AI BS?

How can investors spot AI BS?

3962189473d062fdc76ce9a07cbe89fd?s=128

Shahid N. Shah

October 16, 2020
Tweet

Transcript

  1. AI “washing” allows entrepreneurs and investors to cover up investment

    thesis flaws with hype and “BS”. How can investors spot AI BS? @ShahidNShah Publisher, www.Medigy.com
  2. The simplest answer is: Can the innovator describe the outcome

    of their novelty without saying “ML” or “AI”?
  3. Why? AI is not a product you can buy, it’s

    an experimentation technique which allows for rapidly poking and prodding at huge volumes of previously untapped data to discover facts and relationships about the complexity of our world. Rather than poking and prodding at genes, cells or molecules to see how they interact in labs, “data scientists” use ML and AI to more rapidly discover relationships that would be difficult for humans to “see”.
  4. How to Spot AI BS An entrepreneur that says, "we

    use AI for drug discovery” is just as silly as one who says, “we experiment with molecules for drug discovery.” It does not mean anything. If someone said “you should invest in my company because we know how to culture bacteria" you'd look at them like they were delusional. The same should be done with entrepreneurs who use “AI” and “ML” as if they are the ends, rather than the means to the ends.
  5. Technology has digitized our experiences Last and past decades Digitize

    mathematics & engineering Digitize maps, literature, news Digitize purchasing, social networks Predict crowd behavior This and future decades Digitize biology Digitize chemistry Digitize physics Predict human behavior Gigabytes and petabytes, all sharable Petabytes and exabytes, not shareable
  6. AI must usher in the Scientific Method 3.0 . 1.0

    Identify phenomenon Think about nature Fit to known patterns Guess at answers 3.0 Identify data Generate questions Mine data Answer questions 2.0 Identify problem Ask questions Collect data Answer questions
  7. How will we know if we’ve reached 3.0 ?

  8. 15-year-old student discovers cure for rare disease while gaming Computer

    creates treatment for prostate cancer
  9. Machine Learning and AI in healthcare will be slowed by

    intermediated business models, misunderstood regulations such as HIPAA / FDA QSR and protective regulations such as licensure and credentialing.
  10. Where ML and AI are applicable Therapies Therapeutic Tools Diagnostics

    Diagnostic Tools Patient Administration Payer Admin Clinical Professional Education Public Health Education Patient Education Most Regulation Least Regulation Cohort specific Personalized Risk Data Sharing
  11. Where ML and AI are applicable Therapies Therapeutic Tools Diagnostics

    Diagnostic Tools Patient Administration Payer Admin Clinical Professional Education Public Health Education Patient Education Most Regulation Least Regulation Auto Literature Review Specialty-specific Content
  12. Where ML and AI are applicable Therapies Therapeutic Tools Diagnostics

    Diagnostic Tools Patient Administration Payer Admin Clinical Professional Education Public Health Education Patient Education Most Regulation Least Regulation Auto Adjudication Fraud Detection Quality Compliance Contract Adherence
  13. Where ML and AI are applicable Therapies Therapeutic Tools Diagnostics

    Diagnostic Tools Patient Administration Payer Admin Clinical Professional Education Public Health Education Patient Education Most Regulation Least Regulation Patient Self Diagnostics Unlicensed Pro Diagnostics Digitally and Heuristically Guided Diagnostics Images (self, guided, consulted) Labs and Chemistry (self, guided, consulted) Multi-omics (self, guided, consulted) Molecular Biology
  14. Where ML and AI are applicable Therapies Therapeutic Tools Diagnostics

    Diagnostic Tools Patient Administration Payer Admin Clinical Professional Education Public Health Education Patient Education Most Regulation Least Regulation Auto Triage for Low-risk Augmented Triage for Higher risk Infection control / Anti-microbial Stewardship Consulted Tele Diagnostics Med Device Continuous Diagnostics
  15. Where ML and AI are applicable Therapies Therapeutic Tools Diagnostics

    Diagnostic Tools Patient Administration Payer Admin Clinical Professional Education Public Health Education Patient Education Most Regulation Least Regulation Physical Mental (chat, VR, etc.) Digital (nutritional, etc.) Clinical Research ( “systematic review automation”) Drug Development Clinical Discovery (unattended and digital)
  16. Where ML and AI are applicable Proteomics Genomics Biochemical Imaging

    Behavioral Phenotypics Admin Economics Connectivity Integration Transformation Comprehension Enrichment Insights Cognition No ML or AI possible without these
  17. The State of AI in Life Sciences Find what might

    work Validate in silicon without clinical trials Validate effectiveness in the real world Commercialization Distribute safely and at scale Manufacture safely and at scale Monitor Patients Post Market Improve Safety Post Market Science 3.0 (drug discovery) VERY FAR AWAY AND SPECULATIVE
  18. The State of AI in Life Sciences Find what might

    work Validate in silicon without clinical trials Validate effectiveness in the real world Commercialization Distribute safely and at scale Manufacture safely and at scale Monitor Patients Post Market Improve Safety Post Market Biology simulation VERY FAR AWAY BUT HAS SIGNIFICANT PROMISE, WORTH INVESTING
  19. The State of AI in Life Sciences Find what might

    work Validate in silicon without clinical trials Validate effectiveness in the real world Commercialization Distribute safely and at scale Manufacture safely and at scale Monitor Patients Post Market Improve Safety Post Market Automating clinical trials, detecting fraud TODAY, ACCELERATING, WORTH INVESTING
  20. The State of AI in Life Sciences Find what might

    work Validate in silicon without clinical trials Validate effectiveness in the real world Commercialization Distribute safely and at scale Manufacture safely and at scale Monitor Patients Post Market Improve Safety Post Market Biggest opportunity to learn from other industries TODAY, ACCELERATING, WORTH INVESTING
  21. The State of AI in Life Sciences Find what might

    work Validate in silicon without clinical trials Validate effectiveness in the real world Commercialization Distribute safely and at scale Manufacture safely and at scale Monitor Patients Post Market Improve Safety Post Market Biggest opportunity to learn from other industries TODAY, ACCELERATING, WORTH INVESTING
  22. The State of AI in Life Sciences Find what might

    work Validate in silicon without clinical trials Validate effectiveness in the real world Commercialization Distribute safely and at scale Manufacture safely and at scale Monitor Patients Post Market Improve Safety Post Market The most potential for immediate use (great investment thesis) TODAY, ACCELERATING
  23. The State of AI in Life Sciences Find what might

    work Validate in silicon without clinical trials Validate effectiveness in the real world Commercialization Distribute safely and at scale Manufacture safely and at scale Monitor Patients Post Market Improve Safety Post Market Many opportunities in “real word evidence” TODAY, ACCELERATING
  24. The State of AI in Life Sciences Find what might

    work Validate in silicon without clinical trials Validate effectiveness in the real world Commercialization Distribute safely and at scale Manufacture safely and at scale Monitor Patients Post Market Improve Safety Post Market Once we have lots of data from real-world evidence efforts SOON
  25. What AI is worth investing in? Ask innovators hard questions,

    like: • Which monumental tasks is their novel AI eliminating? • Which significant roles in life sciences or healthcare are no longer necessary because of their novel AI? • Can their AI speed the delivery of patient-facing innovations, improve post-market quality, or speed up regulatory approvals? If you think of AI as a gold rush, pick-axes and shovels are worth investing in if they can be shown to add to the considerable work being done by open source software produced by Microsoft, Google, and others.
  26. How should machines go through medical training? Which medical school

    will have the first machine learning algorithm training department?
  27. Thank You. Join the Innovation Evaluation Revolution at www.Medigy.com Find

    this and many other of my decks at http://www.SpeakerDeck.com/shah How can investors spot AI BS? @ShahidNShah shahid@shah.org