I am not a lawyer, but I’m not surprised if the law lags here. In the past, if an application was accused of being discriminatory, the author can produce the source code that defines why the output is the way it is. But today, the source code doesn’t define the output. The source data does. And AI is designed specifically to be non-deterministic. Meaning that the same input data can create different output data for each person or each request. The whole process is just so different, I suspect it will take time for the legal system to adapt. It’s not clear to me at all what that future system will look like, but I’d love to hear more speculation.
1 Like