AI Now

AI Now

Within the world of data science, there is possibly no field that is as misunderstood as the field of artificial intelligence (AI).  Whether it’s Hollywood films depicting AI robots taking over the world, or sensationalized news stories stating that AI will wipeout the entire workforce, there is a lot of misinformation being spread about this promising field.

The idea that artificial intelligence could take over the world in a dystopian fashion is overblown, but the fear that AI could replace parts of the workforce is a valid concern.  The example of self-driving cars replacing taxi drivers is often cited as an example of an AI phenomena that we will soon be dealing with.  Last week, New York University and the White House co-hosted a series of panels regarding the current state of affairs in the world of AI, exploring its ramifications for social inequality, ethics, health, and labor.  While many experts want to think about the theoretical ramifications of AI in the unforeseeable future, this panel was specifically dedicated to AI developments in the next five to ten years.

The series of lectures opened with an overarching question: How do we responsibly implement artificial intelligence, while accounting for the jobs that will be displaced by AI?

The panelists began by discussing the ability that artificial intelligence has to exacerbate already prevalent economic disparities in society.  AI has the ability to automate work processes, which can lead to unemployment in certain sectors of the economy.  But automation is not a new phenomena.  The 20th century is full of examples of jobs being automated, and how to adjust the workforce, as a way of keeping employment steady.

So what is the answer for automation?  Education.  If replaced workforces can be taught new skills, then the negative effects of AI can be effectively mitigated.

Obviously education is a broad, complicated, and nuanced response that is more easily wished for than accomplished.  And it brings up a broader issue: how should governments be involved in mitigating the proliferation of artificial intelligence?

Our present education systems promise a high school education and, sometimes, a college education.  But what about after high school or college?  There was a consensus from the panelists that there needs to be a set of resources available for adults whose jobs could be replaced by advances in AI.

But artificial intelligence also has the potential to help ease certain tensions, if the data is available.  One of the panelists, Roy L. Austin—the Deputy Assistant to the President for Urban Affairs, Justice and Opportunity—espoused the idea that our data needs to be improved so that AI has the potential to help us gain insight into the injustices prevalent in our society.  AI is only going to work if the data sets it is based on are full and complete.  Austin cited an example of reported hate crimes.  In 2014, there were 759 reported hate crimes in California; in the same year, only one hate crime was reported in the entire state of Mississippi.  There is an obvious disparity in that data, and there was a resounding conclusion that artificial intelligence will only benefit society if our data sets are full and complete.