Artificial intelligence (AI) is based on a set of structures created by people and, unfortunately, as a result of this, biases can creep in unwittingly.
Take the example of the AI created to do facial recognition that was trained on data in Silicon Valley. The data was from mainly white and male faces, because that was the demographic of the majority of people who worked at the tech companies in Silicon Valley at the time. Unfortunately, this was not tested on black faces and so often didn’t work correctly on the greater American public and people around the globe.
AI is being used to automatically do a lot of tasks that were previously done by a person. Things like approving loans or even choosing which candidate, based on a CV, is best for a role. Unfortunately, due to this automation, those people that happen to be the exception to the rule will always be denied, no matter how suitable they might be. On the other hand, people who understand what the AI is looking for can “cheat” the system by saying exactly what is expected in the way it is expected.
We, at Envisage, work with IBM and promote their Watson AIOps (using AI in IT operations) solution. To us, it is important that we provide an ethical solution to our customers.
The AI related to the IBM solution that we sell is discussed here:
https://www.ibm.com/cloud/blog/watson-aiops-bringing-ai-to-it-operations-management
The key sentence that I have found in this blog is: “All AI models are transparent and explainable to earn your trust.” This is very important. Even if our work, which involves monitoring and managing customers' IT environments, doesn’t involve data about people, biases can still find their way in and cause a skewed view of the world.
IBM takes AI ethics very seriously and even has a role called “AI Ethics Global Leader for IBM”. There is an article in the Harvard Business Review that discusses the ethics in AI and IBM’s approach to this situation. The title is: "How IBM is Working Toward a Fairer AI" and it can be read at https://hbr.org/2020/11/how-ibm-is-working-toward-a-fairer-ai.
AI can take a large burden off the workforce by simplifying and automating manual tasks. But it needs to be fair and ethical. If you would like to find out more about IBM’s approach to simplifying your IT operations by using AI (that is fair and ethical), please contact me at bsmythe@envisage.email. Birgit Smythe, Managing Director.
Share