IBM 5 in 5: Unbiased AI will help people trust machines
In the next five years, new solutions will be developed to help control bias in artificial intelligence. AI can be a powerful tool for helping people make decisions— for example, lawyers are using AI to help them conduct legal research and stay up to date with the latest legislation. But when AI is trained using biased data, the bias can be passed into the decision-making process—an issue requiring immediate attention as humans and AI increasingly work together. IBM researchers are developing new solutions to help remove implicit racial, gender, or ideological biases from data sets. Balanced data means balanced decision-making, building a future AI as intelligence we can trust.
via IBMblr = IBM Innovation Culture + Tumblr http://bit.ly/2cbasP7
June 15, 2018 at 08:07PM