Google almost released 100k chest X-ray files, as part of an Artificial Intelligence project. The National Institutes of Health stepped in and advised of the potential violation which halted the exposure of data.
This is another example of how the race to bring AI products to market is putting in jeopardy privacy and security. Google is no worse than any other developer (and in this case moved quickly to end the project and avoid releasing sensitive health data). It is happening across the market. It exemplifies how we are deficient as an industry in establishing broadly accepted ethical AI standards that includes sampling for representative data, proper security controls, checks for potential impacts to life-safety, and especially the generation or exposure of data that can undermine privacy.
AI is a tremendously powerful tool that can bring unimaginable wonders to the world, but it is a tool that can also cause harm and subvert people’s rights. It is important that we recognize the risks and move in unison to establish AI Ethics standards.