Some holiday reading on AI

Fresh in the past few weeks, here are three publications on AI in which I have been involved and which might be of interest.

First, on December 12, my colleagues and I on the AI Index team released the 2018 AI Index annual report. The AI Index, which is housed at Stanford and is a project of the One Hundred Year Study on AI (AI100), draws on expertise and data from all over the world. This year’s report indicates that AI is becoming even more vibrant as a field, based on academic, public interest, and business metrics, and includes many more data points tracking AI around the world.

Second, Martin Ford recently published a series of interviews with leaders in the AI field as a new book, Architects of Intelligence. I was honored to be included along with several of my friends, such as Demis Hassabis, Jeff Dean, Fei-Fei Li, Barbara Grosz, Andrew Ng, Daniela Rus, and many other amazing scientists and technologists.

Third, MGI recently published the latest discussion paper in our Notes from the AI frontier series. Applying AI for social good focuses on the promising potential of AI use for societal benefit, while enumerating some of the current bottlenecks that have so far limited progress towards that potential. A library of about 160 AI social impact use cases that we compiled suggests that existing AI capabilities such as computer vision and natural language processing could be applied to cases in a wide range of domains that touch on all 17 of the United Nations’ sustainable development goals.

Real-life examples of AI are already being applied to a limited extent in about one-third of these use cases, ranging from helping blind people navigate their surroundings to protecting potential victims of sexual exploitation online. Of course, AI is not a silver bullet, but it can be one very useful tool alongside others to tackle these sorts of issues. Scaling up AI usage in order to realize its potential for social good will require addressing key bottlenecks, particularly issues relating to data access and the availability of talent, as well as mitigating risks such as bias and data privacy issues that could harm the very people AI is supposed to help.

As readers of the AI Index report will see, this is an exciting time for AI. I look forward to seeing what the future may hold. Where will more AI soon be deployed for social good? And which of the predictions by leading AI thinkers will start to come true?