Herbert Simon, one the inventors of Artificial Intelligence (AI), once remarked, "Machines will be capable, within 20 years, of doing any work a man can do." Simon made this forward-thinking statement in 1965. What Simon and the other AI proponents failed to appreciate are the numerous unforeseen challenges involved in relying solely on machines.
The AI field first emerged more than half a century ago. But it's only been in the last ten years, due in large part to the rapid growth and democratization of machine learning, that we are starting to see the full potential of AI. Today, there are countless open-source AI tools that help any entrepreneur get up and running on AI quickly, including Google's TensorFlow, Microsoft's Cognitive Toolkit, and Amazon Machine Learning. These advancements have accelerated the transition from a model that involves AI assisting humans to one that involves humans assisting AI. This shift to requiring less human capital creates incredible opportunities for businesses to advance. Yet, many technology evangelists are pushing the needle even further and petitioning the eradication of human involvement.
The media loves to fixate on these advancements in a Jetsons meets doomsday style of rhetoric. If we follow this line of thinking, it seems logical that humans are only a few short steps from being deemed worthless, demoted to the sideline by the superior droids. It often feels impossible to scan any social media feed without seeing robot dogs effortlessly hopping over obstacles or stories of flying cars solving all of our big city traffic woos. The "space race" for the fully autonomous vehicle already dominates tech's daily headlines. What seldom is covered with as much vigor is the fact that, despite the massive acceleration in autonomous advancements, humans are still essential. Maybe more so than ever.
Musk Repeals His Robots
Recently, one of the industry's leaders and biggest proponents of automation, Elon Musk, made a rather important revelation in an interview discussing the production of the Tesla Model 3. The backstory is that Musk's two assembly lines were not churning out cars fast enough and Wall Street was not happy with the fact that Tesla was hemorrhaging cash. According to Bloomberg, Tesla burned $3.48 billion in just over 12 months, equivalent to $6,500 per minute.
Tesla's original assembly lines were coined the "alien dreadnought" and were rife with robots piecing together the electric vehicles. But the robots kept breaking down causing production to stall and pushing Tesla's Model 3 goals further and further out of reach. Human workers began needing to rip out conveyor belts and carry parts by hand to keep production moving.
Musk was forced to halt production and face the reality, acknowledging, "Excessive automation at Tesla was a mistake." Musk has since removed the Model 3's fully autonomous production line and replaced the lines of robots with people. As Musk toured the new assembly line for an interview with 60 Minutes, the stark difference from the doted droid autonomy was readily apparent. One of the workers remarked that humans are just better at dealing with unexpected circumstances than robots.
The balance between human intervention and automation still seems poorly understood and rife with challenges. It isn't just Tesla and its assembly line feeling the burn. Everything from rouge AI-driven chatbots to societal alteration algorithms are glaring reminders of the dangers of artificial intelligence leaping too far too fast. One of the most interesting eye-opening books I have read in the last year is Cathy O'Neil's, Weapons of Math Destruction.
O'Neil analyzes how the algorithms designed to help us weed through the copious treasure trove of big data can be more dubious and hazardous if left to run on their own. These solutions that are deemed "highly scalable" can quickly reinforce and amplify biases leaving society worse off with little recourse.
The exciting world of fully autonomous vehicles has also instigated a debate on the moral ethicacy involved with a machine making all the decisions instead of the driver. If a human walks in front of your moving autonomous vehicle, should the car protect you over the other human? What if it was a child? Is society ready to program vehicles with moral decisions or do humans still need both hands on the steering wheel in an AI world?
Those trying to push their advancements in AI forward as fast as possible need to, not only think of the accolades involved in leaping but also the repercussions of recoil. As for Musk, aside from how Wall Street will inevitably value his publicly traded assets and the mounting HR nightmares that still need to be sorted out, Musk remains one of our generation's most ingenious and avant-garde inventors who continue to push innovation forward. Yet, not every tectonic shift forward is the right one and, as Musk stated when realizing the limitation of a fully automated process, "humans are underrated."
See the entire 60 Minutes episode here: