Trustworthy AI

Chapter 8Privacy

BAM Inc.'s Chief Data Officer Marguerite could not have been more frustrated. Aftermarket product data flows should not have been diverting her attention. Nevertheless, she was almost blind to how products were functioning in a particular country whose domestic laws placed heavy restrictions on how data could be shared and exported beyond the national borders.

From her perspective, the laws made no sense. Their products featured edge computing capabilities feeding highly detailed data into a data lake where several AI systems could monitor function in almost real time. In terms of product viability and durability, BAM Inc.'s customers were getting frequent notices on how to maintain their products and potentially replace them when the time came. It was a mark of quality.

But due to domestic privacy rules, that one country just would not release the data she needed to maintain BAM Inc.'s quality standards (to say nothing of improving future products). She could not afford to ignore the problem, but the path forward was unclear.

When it comes to privacy in AI, laws and regulations are the result of where human stakeholders place ethical value, be they government bodies, consumers, or others. These can be divergent across geographies, and every company using AI must contend with the patchwork of privacy requirements wherever they operate.

The issue of privacy is most visceral when it comes to data about people. Humans create troves of data points that reflect ...