In an interview, Jennifer Shanker, journalist, talks about the pursuit for ethics in AI. She talks about the need for machines to pattern match in order to solve problems that humans have not yet been able to solve. The uncomfortable points for the general population arise when technologies like autonomous vehicles are concerned. People are concerned about the decision making process that is needed in certain edge cases of a product’s functionality. Leaders in technology agree that there is a lack of ethical governance in the AI domain.
Shanker interviewed a researcher who researches ethical encodings. Her specialty is decision support systems, where she works together with professionals from other disciplines. He niche is interesting, because she reattaches ways to create technology to aide group decision making processes. Franchesca talked about there being no universal principle to ethics in computing. It seems that ethical issues are very task specific in implementation. Since there is no agreement with human values, it is challenging to succumb to a set of agreed upon universally acceptable rules.
To create ethical APIS, there are two approaches that were mentioned, one is rule based approached. Where the machine follows a set of rules. This approach follows a top to bottom approach. The other is a bottom up approach that allows the machine to observe learning that will allow the machines to learn from behavior it sees. Both have shortcomings, and there needs to be a hybrid model that uses both approaches.
The overall sentiment was that modern technological dilemmas need to be addressed by machines in ways that make sense. It was interesting to me to listen that machines compliment human in their decision making process. It not that humans or machines are better, but both together brings efficiency. It is interesting that the future of an AI led world will ultimately be decided by humans.
We need to decide if we want a future that is led by artificial intelligence.