Anticipating potential risks

Part of the specialization that takes place in the current form of an AI PM is knowledge of what can go wrong. Your AI features may be taking certain liberties with your customers’ data, and it’s your job as a PM to anticipate the potential risks your product might create for your users. As a PM, you’re looking to address the most important business problems your users have with your product, but you’re also looking to minimize potential adverse effects.

I know, we got you all jazzed up about being a fearless AI product manager only to then bring up the things you should be afraid of because we want to make a distinction. I’d rather you be afraid of the potentially harmful side of AI than be afraid of the complexity of AI systems. Understanding the structure and complexity is the easy part. Making sure this complexity is handled properly is the hard part. For example, rather than worrying about where to start with...