AI and Medicine
What are the criteria for an effective medical AI?
Despite the early struggles of establishing AI in medicine, the more recent implementation of DL looks promising. However, there are things we can learn from these past failures to help build robust criteria for further AI applications to benefit the medical profession:
1. Capable – To have any utility in medicine, AI needs to be at least as accurate as a human equivalent in fulfilling its purpose.
2. Practical – The systems need to work efficiently and ideally should be able to integrate with pre- existing infrastructure, as well as requiring little maintenance or training to operate.
3. Affordable – The technology needs to be reasonably priced to be a widely used tool that minimizes healthcare inequality rather than exacerbating it.
4. Ethical – Much of the work surrounding AI involves decision-making based on patient data. Ensuring that the data used is secure and produces a fair, unbiased and accurate assessment is of the utmost importance. Following this, the establishment of a legal framework of responsibility will also become increasingly relevant as medical AI becomes a reality.
If AI is to have a place in the future of medicine it must have some form of response to each of these criteria. There are several different considerations to be made for each of these points and so for this dissertation, the following perspectives will be discussed.
(1) How can this medical AI be integrated seamlessly with existing medical equipment, systems and workflows?
(2) How can the quality of data can be such to ensure that the decisions AI makes are robust and reliable?
(3) With the possibility of personalized medicine, how does AI affect the patient experience, and doctor-patient interactions? What is the impact of this technology on the healthcare workforce?
(4) What are the ethical considerations that arise as a result of introducing a technology as novel and disruptive as AI. How might cyber security protect the constant flow of patient data used by AI? How to protect against the biases and discrimination AI could reinforce and, perhaps most importantly, who holds responsibility should an AI fail in its task? By the end of this dissertation, we will conclude that not only is it likely that AI will have a role in the future of medicine, but that it has the potential to revolutionize healthcare and improve the patient experience, particularly in those systems that are under immense pressure.
35
Made with FlippingBook - PDF hosting