Computational linguist Frank Rudzicz begins his lecture by drawing an analogy between automated news and automated healthcare.
Facebook’s algorithm learns from your behavior, identifying the news sources and headlines you interact with and presenting you with content that it deems fit within your interests.
The future of healthcare may be automated in a similar way: machines learn from health data and present solutions in response to that data. And there are risks involved.
The threat, Rudzicz says, is not “disobedient AI,” but rather obedient AI with “lazily-defined objectives.” Social media has shown us that algorithms designed for seemingly harmless purposes can have “disastrous,” if unintended, consequences when mixed with “hidden human biases,” Rudzicz states.
In this lecture, Rudzicz also identifies the current trends in AI:
- Deep neural networks
- Big data
- Recurrent neural networks for temporal, dynamic data
- Reinforcement learning
- Active learning
- Telehealth and remote monitoring
- Causal, explainable models
Watch the full video to learn about each of the seven trends listed, and visit C4E for more information on the ethics of AI.