Human in the loop

The integration of AI into healthcare represents one of the most promising avenues for improving quality of life on a global scale, while simultaneously presenting some of the most pressing ethical challenges. These include ensuring that AI systems, their training data, and their predicted outcomes are used equitably and contribute to positive health outcomes for all populations.

The following presentation by Conor Judge, “The future of AI in medicine” for TEDx Talks consolidates our discussions from AI and Healthcare, with a focus on prioritising human compassion and understanding.

TASK: As the final task of this section, choose three of the following questions and discuss them with a human. No chatbots this time. Enjoy your human-human interaction.

1. What simple, explainable principles should guide the development of AI tools in healthcare, especially when human lives are at stake?
2. Should AI systems be allowed to make clinical decisions without human oversight? Under what circumstances?
3. Who is responsible when an AI system makes a harmful or incorrect decision in a clinical setting?
4. What steps can be taken to ensure that AI systems do not perpetuate or exacerbate existing health disparities?
5. Is it ethical to train medical AI systems predominantly on data from high-resource settings and then deploy them globally?
6. How can AI be designed to augment, rather than replace, the human–human relationship?
7. How do we ensure that AI tools are accessible and beneficial to populations with limited digital infrastructure or medical records?
8. Who is ultimately responsible for a clinical outcome when decisions are made jointly by a human and an AI system?
9. Are there any no-go zones for AI in healthcare? If so, which and why?
10. What source of bias is more significant in healthcare, that of the human, or that of an AI system?