On Artificial Intelligence and Ethics
By David Tamayo, CIO, DCS Corporation
A day doesn’t go by without the words “artificial intelligence (AI)” appearing in mainstream media referencing everything from self-driving cars (robots) to automation galore in our “smart” things. Yet, we rarely ever see the words “ethics” and “morals” associated with AI. Luminaries such as Stephen Hawking, Bill Gates, and Elon Musk are now warning about the risks of the rising robots against the falling apes (yes, I’m talking about us humans). These risks come not just in the loss of millions of highly paid jobs, but also from the potential harm that might come from automating critical infrastructure (e.g. electric grid, water purification, etc.), our Nation’s military, our markets, and the general outsourcing of our brains to AI.
Let me ask: Would you create, support, and help develop a technology that will destroy your kids’ future? Some claim that is exactly what we are doing with AI. Yet not doing AI would kill your own job as your company loses its edge against competitors. For example: Uber, Tesla, Volvo, and others have demonstrated their ability and willingness to deliver self-driving trucks over the next 3-5 years. Unlike humans, an AI truck will work 24/7 covering more territory for less cost than any single human possibly could, and it would avoid most of the 4,000/year truck-related fatalities as well as 100,000 injuries. That is a great thing! Yes, but on the other hand, in a relatively short period of time it will deprive 1.7 million people from making a living driving. As they say, if you are not disrupting, you will be disrupted.
In 2014, roughly the same revenue was generated by the top 3 companies in Detroit as did the top 3 companies in Silicon Valley, yet the companies in Silicon Valley had 10 times fewer employees. Fewer employees translate into less payroll taxes (i.e. automation pays far less taxes than people) and less expenses, which sounds great. Yet this also creates greater inequality in the population by increasing extremely high profits to corporate owners at the expense of human employees. With the increasing ability of AI chat-bots to model human conversation and relationships, they started to replace even the “speaking” jobs. Many of these bots, under manipulation from a foreign power, are already suspected of having influenced last year’s U.S. elections.
As more things become “intelligent” and as Artificial General Intelligence becomes a reality, the number of people in the world whose jobs will evaporate will be faster than ever
Click-bait headlines, search engine results, GPS info, photographic meta-data, and similar metrics are now allowing corporate AI systems to know you better than you know yourself! Keep in mind that for much of the “free” social media used today, YOU are their product!
There are lots of other ethical and moral risks that rarely get any attention. It turns out that AI becomes smart in much the same way a child becomes smart: by learning. As with children, sometimes AI will draw the wrong conclusions. For example, an AI system used by the courts in Florida to determine the probability that a defendant would break the law in the future was recently found to be racist, encouraging the courts to hand out harsher sentences to minorities. This should not be surprising since these systems learn from our biased society, which is infused with biased historical data. In other words, the old adage still applies: garbage in, garbage out.
Ambiguous instructions could create a disaster for AI’s. Consider the robotic command “make your owner as happy as possible”. After some thought, the robot gives its owner a lobotomy leaving him in a state of permanent bliss, sitting in a corner, drool dripping from his mouth. Self-driving cars have a hard time dealing with a bag blowing across the road: is it a deer? Is it smoke? Is it solid? Such mistakes can easily kill a few people. Now consider a similar level of Artificial Stupidity relating to nuclear weapons… such error could destroy all humanity. It is no secret that armed drones are currently flying the skies of many countries and AI is, with some human supervision, identifying the enemy and civilians, deciding who lives and who dies. Now imagine these systems in light of the recent Equifax and Uber hackings, Wannacry, and Wikileaks’ NSA/CIA hacking tools.
The main risk with AI comes in the form of unintended consequences… asked to eliminate cancer, the AI kills every person with cancer – problem solved! As more things become “intelligent” and as Artificial General Intelligence becomes a reality, the number of people in the world whose jobs will evaporate will be faster than ever. For the first time in history our brains are being replaced instead of our muscles – so if both brains and muscles are replaced, what is there left for humans?
Imagine an AI system trained to instantly recall, interpret, and argue every doctrinal, religious, philosophical and theological argument of a particular religion. Such theologian would be better than any priest, pastor, rabbi, imam or any other religious expert. Philosophy and theology will have to deal with questions like: would these just-as-smart-as-human AI be endowed with a “soul”? Should they proselytize as required by certain faiths? Should these AI’s pray? If you think these ideas are just too crazy, ruminate in the fact that a few weeks ago Saudi Arabia was the first country in the world to give citizenship to a robot named “Sophia”. This prompted many critics to note that now a robot has more rights than migrant workers and women!
Can all this AI progress be stopped / slowed? Where will it end? Know that the issue is not AI. NO GOVERNMENT POLICY anywhere is being created to deal with the many issues these new potential “AI persons” will pose very soon. AI and the ethics/morals associated with them are being driven by private industry based on potentially amazing profits, so what could possibly go wrong? These and many other dangers hover over AI’s progress, yet AI technology is “very cool” so most of us just gaze in owe; we are like a moth flying straight into the very nice, beautifully bright, warm, and gorgeous flame!
Applications of Big Data and Predictive Modeling
Somesh Nigam, Ph.D., SVP and Chief Data Officer, Blue Cross and Blue Shield of Louisiana
Grow and Supercharge with Artificial Intelligence
Deep Varma, VP of Engineering, Trulia