A: KATHERINE ELKINS | Associate Professor of Comparative Literature and Humanities JON CHUN | Visiting Instructor of Humanities, Affiliated Scholar in Scientific Computing
For most of human history, automation has dramatically increased our material comfort, wealth and well-being, freeing humans from the “3 Ds” of “dull, dirty and dangerous” work while providing new and more stimulating work opportunities. The majority of us believe our jobs will remain immune, and that automation will augment rather than replace human labor.
But recently, anxiety about automation has increased. Automation is suspected of playing a role in the stagnant wage growth and underemployment that has plagued developed economies for the past half century. Artificial Intelligence also is demonstrating the potential to automate almost every human activity.
The newest wave of AI-powered automation is less about automating simple physical labor and more about empowering automation with human-like perception, communication and cognition. Studies have found that up to half of existing jobs are at risk for automation within two decades, and it’s possible AI will outperform humans in all work tasks within 45 years.
Still, some “dull and dirty” jobs pose real challenges to automation. Ironically, AI turns out to be better at manipulating abstract symbols, exploring innovative designs, and sensing human emotions than cleaning toilets. Imagine a future dystopia in which we clean toilets while AI writes symphonies, serializes TV shows and crunches numbers to make life’s important decisions.
What can we do as a society if many of us cannot compete with automation for traditional jobs and new job creation fails to keep pace? Small scale experiments in Universal Basic Income are testing it as a viable alternative to a work-based economy. Taxing robots in proportion to the human labor they replace may preserve government revenue and slow its adoption. Congress has banned self-driving trucks, and human oversight is mandated for data-driven criminal sentencing and medical diagnosis.
Still, there remains the question as to how individuals will reimagine their purpose and identity without work. Automation may force us to rethink the fundamental premises of our economy, laws and ethics as it accelerates wealth concentration, creates vast power imbalances and swiftly outmaneuvers any regulations designed to control it.
We all need to ask and answer these big questions to ensure that an automated world is still one we wish to inhabit. Only in doing so can we determine how to maintain a humanist world in the face of an increasing post-humanist onslaught.
KATHERINE ELKINS | Associate Professor of Comparative Literature and Humanities JON CHUN | Visiting Instructor of Humanities, Affiliated Scholar in Scientific Computing
When IBM Deep Blue defeated chess champion Garry Kasparov in 1997, The New York Times reported it would take 100 years before a computer could defeat a human Go master. Go is a vastly more complex game than chess that requires computers to rely on heuristic shortcuts that mimic human intuition.
Yet it was just 20 years later when AlphaGo twice defeated Lee Sedol, an 18-time world Go champion, and became the first computer to beat a top-ranked human in a Go match.
Astonishing advances like this have led to a global “arms race” in artificial intelligence, as companies compete to acquire top AI talent. Last year, China announced a multibillion-dollar initiative to become the world leader in AI and, recently, Russian President Vladimir Putin proclaimed, “Whoever becomes the leader in this sphere will become the ruler of the world.”
While AI experts work to develop smarter AI applications for all of us — from driverless cars to personal assistants — fewer have taken up the broader challenge to ensure we don’t become, in Henry David Thoreau’s words, “the tools of our tools.”
What we need now are humanists conversant in AI who can critique and shape the future that AI may restructure. After all, AI forces us to ask questions about what it means to be human. And answering these questions will, in the end, be more important than AI milestones like AlphaGo. The only way to answer these questions is to develop an understanding of the world that is both broad and deep, since these questions cannot be answered within any single discipline or major.
No one in 1997 could have predicted the advances in big data, computational power and algorithms that are making AI increasingly powerful and inexpensive. How, then, can we predict what AI will look like 20 years from now? Even the experts are poor at forecasting this future. But the rapid and revolutionary changes being brought on by AI compel us to continue putting the human at the center of our technological world.