Workplace Harassment Reporting—Is There an App for That?
the situation
According to the EEOC and other sources, a large majority of all workplace harassment goes unreported. Perhaps this is because employees don’t feel comfortable reporting workplace harassment. But perhaps it is also because employers don’t always ask the right questions or utilize the best information-gathering methods. So how can artificial intelligence be used to address this issue?
the ruling
A number of companies are working on developing artificial intelligence specifically designed to target workplace harassment. One such company has developed an app called Spot. Spot is a chatbot and the idea is that it asks open-ended, “cognitive interview” style questions, much like those used in law enforcement investigations.
If an employee is subjected to what he or she believes to be harassment, he or she can go to Spot and talk to a chatbot through a text-message type of exchange. The chatbot asks some questions about what happened and then uses natural language processing to ask follow up questions about certain people and places that were mentioned by the employee. Spot then puts the results together in an encrypted report and the employee can either download it and send it to himself or herself, or send it to his or her employer. If sent on to the employer, the reports can also be sent anonymously.
the point
The makers of Spot and those working on similar AI products are hopeful that the availability of this type of app will make employees more likely to report harassment. But they also believe it will result in more accurate reports—after all, theoretically, the chatbot won’t have any bias that could affect the way questions are asked and is less likely to miss key details.
On the flip side, there are potential security issues that could arise out of the use of such an app. And then, of course, employers will be charged with having knowledge of potential harassing activity and thus will need to be prepared to investigate and take action when necessary.
