PORTLAND, Maine — Artificial intelligence has crept into your life in ways you cannot imagine, ways that go far beyond Netflix’s knowledge of what types of movies you like or Amazon’s ability to predict what books you might purchase. It does amazing things—but as 207’s tech guy, Rich Brooks of Flyte New Media, points out, AI also has a dark side. How concerned should we be? Here are the talking points Rich laid out:
We should absolutely be aware of the real and potential problems with AI. These aren't concerns from the tinfoil hat brigade. Stephen Hawking said that AI's impact could be cataclysmic unless its rollout is strictly and ethically controlled. Elon Musk called AI much more dangerous than nukes. Some of the smartest and most forward-thinking people around are concerned about the dark side of AI so it's worth examining.
You mentioned "real and potential" problems with AI. What are some of the current, real problems?
One big, immediate concern is job loss. Organizations like the Brookings Institute have forecast that 36 million jobs could be at risk because they're made up of highly repetitive tasks, meaning a machine with AI could do it. And white collar jobs aren't exempt from automation, either. And as robots become more dexterous, they can take on more tasks.
One caveat here is that since the industrial revolution and even before, certain jobs have always been at risk of going away because of technological advancements. Usually new jobs are created because of advancement. Whether this is true with AI remains to be seen.
Beyond job loss, what other problems exist today?
AI is a reflection of the programmers who create it. Most AI programmers are white, male, come from high socio-economic backgrounds, aren't handicapped, and so on. They have a certain world view, which either intentionally or unintentionally, goes into the programming.
This may not seem like a big deal, but there are plenty of examples of AI having racist biases, including a Google AI program that put photos of Black faces into a folder entitled gorillas or another AI that couldn't identify the face of a Black researcher until she put on a white mask.
There are also privacy concerns. With facial recognition, a powerful tool in the AI toolbox, law enforcement is able to track people without their permission. While there are some laws being enacted to protect citizens against that, you can be sure that totalitarian countries will have no problem leveraging these tools. Or that fears around terrorism may allow mass facial recognition of the public to occur.
And looking forward, what potential problems exist?
We're already seeing AI being used in weapons, and the idea of a future war fought with AI is only a matter of time. With autonomous weaponized drones, we're putting military decisions into the hands of, well, something without hands or a heart. What if AI decides to launch a nuke or chemical weapons because that's the optimized outcome? Even if you don't believe the US government would rely solely on this technology, could you say the same about every government? Or every anti-government organization?
What can people do?
A lot of this is happening in the private sector, where profit is often the number one driver. However, there is still government oversight and regulations. Reaching out to your representative may be one way let your opinion be known and fight the power.