Research in AI-safety seems to be going in many directions. Some scientists focus on containing AI, while others talk about building ethically and cognitively superhuman AIs. This article is a good reminder of how difficult it is to represent rules or statements correctly. Building ethical or legal statutes into knowledge bases or bias systems will be challenging to formulate. At this point, implementation using fuzzy and nonmonotonic logical systems is not exactly a walk in the park. Using #machinelearning might yield many surprises as well. Jaan #Tallinn and his colleagues seem to be very cautious at this point. Results produced by the Future of Humanity Institute in Oxford will be significant.
#FHI #AI #AIsafety #knowledgebase
https://www.bbc.com/worklife/article/20180723-the-commas-that-cost-companies-millions
No comments:
Post a Comment
Note: Only a member of this blog may post a comment.