Good afternoon and happy Friday! Artificial intelligence might be the hottest topic of 2019, as we move toward applying AI to augment the human experience. The only question is how do we apply it in a way that makes sense for us, and a way that is ethical? Here are five essential characteristics of ethical AI:
Fairness and Inclusiveness: AI systems are new and exciting, kind of like a child that is ready to learn about just about anything. Learning machines should be taught to treat everyone fairly, because the purpose of this technology is to empower and engage everyone. If AI is to work for all users then we need a lot of data from many perspectives to help build create a fair unbiased machine. In addition the people that are teaching a learning machine need to be from many different backgrounds to share different inputs to avoid bias.
Reliability and Safety: Artificial Intelligence is continuously developing and updating its knowledge, yet there is still an abundance of unexplored applications for the technology. That is why it is important to monitor how safe and reliable it is to use this technology. The primary focus should be to speed up tedious tasks to allow for individuals to become more efficient and effective in their work. AI should enhance and augment people’s jobs, and allow them to focus on more important tasks.
Privacy and Security: Like any revolutionary technology advancement there is huge potential for what it can bring to societies and the positive impact that it could have on the future. However there are also privacy threats where information can go into the wrong hands. By using AI systems we must ensure that data is securely stored, and the privacy of individuals and companies alike is respected.
Transparency: AI is being used for many different reasons by a wide variety of people, it is important that all users understand how AI does the work it is programmed to do. Understanding how AI makes its decisions gives individuals the insight needed to troubleshoot and manually fix issues efficiently and effectively. It also allows them to assess if the AI made the right choice and train the model to be more accurate in the future. This would allow us to have a way to follow what the machine is doing, and give us the choice to override certain decisions that we find to be inaccurate or unethical. By understanding what a machine is doing and how its deriving its solutions we begin to trust it and hand it deeper more intensive problems to figure out.
Accountability: AI needs to have the ability stop and alert the necessary individuals in the case of any errors that are manually caught. By being able to hold AI creators accountable it enforces the partnership individuals have with it, we want to understand why things are happening and not just hear “because that’s what the machine says”. By holding the creators of the machine accountable in situations like a medical surgery where a doctor can choose to follow his own procedure and “pull the plug” on what the AI is suggesting in order to continue safely.