Building a safer, smarter AI starts with learning from the past and identifying potential risks. In order to mitigate these risks, a diverse knowledge-sharing community must be created.