First name" is a proxy for gender and can be a proxy for ethnic origin or country of origin. As a result, the team removed 'name' as one of the variables in the model. 3. Put yourself in the shoes of the people the technology may affect Since none of us can fully understand someone else's reality, AI developers should seek direct feedback from customers. This way they can find out how their technology affects a certain group of people. For example, at Salesforce, our Research and Insights teams host User Summits.
Employees and consumers then come together to gain a good insight into the context, needs and concerns of both groups regarding data collection and personalization, among other things. This allows us to develop features and guidelines for our products that everyone involved can trust. 4. Don't fall back on the Israel phone number list motto 'move fast and break things' Honesty is a complex concept that requires quite a bit of effort. Especially when AI impacts human rights (justice, privacy, health, access to food, or housing), you are even more responsible for an accurate, fair decision. While the industry has collectively moved away from the “move fast and break things” mentality in recent years, there is a chance that we will take a step back due to COVID-19. Organizations try to react quickly to adapt.
But during a crisis, you can't react quickly at any cost. The strength, and the risk, is that AI can make predictions about people and their behavior at lightning speed and on a large scale. have learned the hard way that they are slowed down if they don't look at possible negative consequences. This can cause a lot of damage to the process. If you build AI right away in a responsible way, you can focus on making the world a better place. You don't have to run after everything you've broken. Now more than ever, spend time on thoughtful, productive friction to ensure accuracy and fairness.