Ms. Hao is the artificial intelligence reporter at MIT Technology Review where she helps to demystify AI through her semi-weekly newsletter The Algorithm. Karen is formerly a data scientist and reporter for Quartz where she constructed machine learning models, built chatbots and covered the future of cities. Before that, she was an application engineer at the first spin off of Google X.
Karen is a both a journalist and engineer, working to create an economically vibrant, socially inclusive, healthy and sustainable future. She very much enjoys operating where storytelling and technology intersect. Her writing has also appeared in Mother Jones, Sierra, Grist, How We Get To Next, New Republic, and other publications.
Massachusetts Institute of Technology, (B.S.) Mechanical Engineering
“It amazes me when people think numbers and math are somehow impervious to inaccuracy, bias, manipulation. Algorithmic bias is very real.”
Upcoming Speaking Engagements
Data Ethics: Exploring Vice and Virtue in Big Data
SuperBot 2019: The Premier Conference for AI, Chatbots, and Voice Assistants
Karen Hao on Artificial Intelligence
Have you ever tried explaining what artificial intelligence is to someone? The field is continually advancing and thus our idea of what constitutes AI often gets distorted. We use the term so much (especially in marketing) that it has become difficult to separate fact from fantasy.
Fear not because this flowchart shows you how to determine if something is indeed using artificial intelligence.
The term “machine learning” is bound to come up sooner or later in any conversation concerning AI in marketing. Again, it’s one of those terms that marketing executives and sales like to use, even if they don’t quite understand what the phrase means.
Next time someone tells you they’re using machine learning, use this flowchart to determine what type of learning they’re really using.
Have we reached peak AI? Karen Hao’s study of the last 25 years of artificial intelligence research suggest that the end is near for the deep learning era.
The quest to create intelligence is a notoriously difficult problem to solve. In the ’60s we tried neural networks and in the ’70s we switched to symbolic approaches. Knowledge-based systems were popularized in the ’80s but soon became unwieldy. The use of Bayesian networks was explored in the ’90s until support vector machine gained favor in the ’00s. Neural networks, through the use of deep learning, have made a comeback in the present decade.
However, this era could come to an end as the AI research community develops more sophisticated capabilities to replicate intelligence.
Remember to have a look at our list of who is who in AI marketing.