The NHS, Centre for Data Ethics, and Nuffield Council have put together a code of conduct for AI use in health care.

My question is whether item 7 as follows is possible and in what ways might it be:

7.1 Demonstrate the learning methodology of the algorithm being built.

7.2 Aim to show in a clear and transparent way how outcomes are validated.

  • 2
    $\begingroup$ By possible you mean Aim to show in a clear and transparent way how outcomes are validated.? $\endgroup$ – DuttaA Feb 20 '19 at 10:17
  • $\begingroup$ indeed. How to show how things are validated in a clear way... Or I guess: if we do show then maybe this showing will demonstrate how insufficient this probably is to validate proper use of the algorithm $\endgroup$ – benbyford Feb 20 '19 at 11:08

The Artificial Intelligence Code of conduct of the UK NHS was written for practical purpose. It defines, what hardware and software in the wild have to do and what not. The idea is to improve the workflow in the hospitals and make healthcare better. To interpret it right, we have to take a look at the gap between healthcare in reality and healthcare in AI research. In reality, no kind of Artificial Intelligence is used. Most hospitals have only a small amount of computers. The reason is, that all the new technology developed in universities can't be adapted to practical usage. That means, on the one hand, the current research has developed medical nanorobots, and at the same time real hospitals are using Windows XP plus MS-Word for writing their letters and print them out on paper.

The probability is high, that the gap between AI in research and AI in real hospital will increase in the future. And if in reality no advanced technology is used, it is easy to regulate it and lower the risks. In real healthcare system there are no kind of algorithms available. Robotics and big data is located outside of hospitals. To answer the question direct: yes it is possible to fulfill the code of conduct of the NHS. Because the danger of technology which isn't available is low. If no robot is used for treating patients, the robot won't hurt them. And if the hospital has no database with all the personal healthcare data of the patient, the database can't get lost during a hacker attack.

|improve this answer|||||

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.