Ethics — the next frontier for artificial intelligence
This post was originally featured on TechCrunch.com
AI’s next frontier requires ethics built through policy. Will Donald Trump deliver?
With one foot in its science fiction past and the other in the new frontier of science and tech innovations, AI occupies a unique place in our cultural imagination. Will we live into a future where machines are as intelligent — or frighteningly, more so — than humans? We have already witnessed AI predict the outcome of the latest U.S. presidential election when many policy wonks failed.
Perhaps we are further along than we thought.
In October, then-President Obama hosted the White House Frontiers Conference, which focused on the leading global technologies featured in the November issue of WIRED, which was guest edited by Obama. Given that our country was founded by innovators and disruptors who envisioned and executed exciting new technologies of their time — like the postal service, the precursor to our inbox struggles — it feels like we’re coming full circle to have our then-current president comment on the next wave of technology that will keep the U.S. on the forefront of innovation. Artificial intelligence is at the heart of that innovation.
Of course, now that the 2016 election has come to pass, there’s an elephant in the room: Donald Trump. What President Trump will have to add to this conversation is, as of now, another great mystery.
Though many might argue the point, Obama put it to WIRED that now is the best time to be alive. The next four years will admittedly be different, but the point stands. There are no real reasons to believe technology’s rapid pace will slow much, and the same goes for AI.
So far, we’ve seen just the tip of the AI iceberg through technology such as virtual personal assistants, self-driving cars and credit card fraud prediction technology. If we pause for a second to recognize how incredible it is that we can ask our phones for directions, then sit back as our self-driving Uber takes the wheel, as well as get an email update instantly when our credit card may have been hacked based on algorithmic learning, then we might just feel like our current reality is a sci-fi plotline.
The implications of AI will be important moving forward, and may require more attention on a federal level. In a report released by the White House on the current and future state of AI, leading innovators considered not only the technology that will drive AI, but also the ethical considerations that must fuel its growth. How should we regulate automated cars to ensure public safety? How can AI be used to streamline government operations and provide new jobs (for humans)?
No other technology has the kind of far-reaching, global implications that artificial intelligence does across so many industries, from healthcare to transportation. Which is what makes it so fascinating. For a technology designed to mimic human-like intelligence, artificial intelligence captures our collective imagination in a way no other technology ever has.
Machines have already surpassed humans in terms of image recognition ability. In the next 20 years, experts predict that machine learning will continue to make great strides on a number of human tasks. If this innovation is done in an ethical way, we can build a future in which humans are not competing with machines or being overtaken by robots, but instead entering into a new era of collaboration that frees up the human spirit for more meaningful tasks that require emotional intelligence.
This is where public policy must keep pace with the rapid advances happening in AI technology. It is not unreasonable to hold Donald Trump accountable for ensuring such policies are protective of both private and public interest.
Uber recently debuted a fleet of self-driving cars in Pittsburgh. As mind-boggling as hailing an automated taxi at the tap of a smartphone screen is, imagine the mental traffic jam caused by trying to regulate transportation and safety laws to accommodate for this next wave of driverless cars. The DOT has their work cut out for them, as too will all of our elected officials.
But the technology AI has already brought to life is still narrow in focus compared to its far-reaching potential. Scientists call this narrow versus general AI, and while we’re still living in the age of the former, it’s only a matter of time before we achieve the latter: an age in which machines can function across a wide array of tasks as intelligently as a human. Voice-command personal assistants and driverless cars are pretty cool, but they’re like riding a bike with training wheels. We’ve only just begun.
This is just the type of problem that makes AI a uniquely American challenge in so many ways. As we create new technologies that will alter the fabric of our daily lives, we must simultaneously implement the policy solutions that will protect AI from going the way of sci-fi movies — in which machine learning approaches the singularity and humans are trumped by the very machines they’ve created. We must take a democratic approach to future technology. We have to ensure that science is aligned with the public good.
Donald Trump, then, will have to somehow reckon with the inevitability of automation, even as he promises to bring back manufacturing jobs. Not everyone has high hopes; in fact, 100 technology leaders pled in an open letter that a Trump presidency would be a “disaster for innovation.” That one-fourth of entrepreneurs are immigrants, and tech companies are apt to recruit talent overseas, could bolster this theory considering Trump’s harsh stance on immigration. Nor does it bode well that Trump is likely to put national security interests over those of privacy.
But fear not. Given that sectors like technology move globally and have power beyond policy, their path won’t much be hindered by a Trump presidency. Even better, advances in artificial intelligence are already supporting the notion of a future in which machines strengthen human perception, rather than deplete us of it. For instance, one study cited in the White House report surveyed a radiology experiment in which images of lymph nodes were displayed to AI technology and a radiologist to determine whether they were cancerous. When AI and the radiologist pooled their knowledge and approached the task as a team, the error rate dropped to 0.5 percent. This is an 85 percent reduction in error! Studies like this make an undeniable case for the good side of AI.
As for preventing the sci-fi side of AI? That’s where policy must intervene, to rise to the challenges of machine learning. Though Trump has no clear policies on the matter as of now, we can only hope technological democracy succeeds as we move forward, whether because of or in spite of American leadership. We may find that while his administration is less concerned about “ethics,” per se, the prospect of a harmoniously automated future is a deal too good not to secure.