Learning and Artificial Intelligence
The media is chock-full of harrowing narratives about how artificial intelligence (AI) can go wrong. If the science fiction genre is to be believed, we’re just a few lines of code away from world-ending rogue AI. A few mistakes away from a singularity which will make humankind entirely obsolete.
Perhaps in an effort to seize upon this anxiety, many media outlets are spreading their own narrative of fear, uncertainty, and doubt. And when they’re not ranting about the singularity, many journalists talk at length about the raw power of the technology. The notion that with AI, you can easily revolutionize your industry and propel your business light-years ahead of the competition.
AI and its close companion machine learning are incredibly formidable technologies. That much is indisputable. They’re already changing the world in new and unexpected ways, from virtual assistants like Siri and Alexa to fully self-driving cars.
At the same time, I’d like to clear the air about what AI is and isn’t. Because it’s neither the holy grail that people believe it to be nor the catastrophic invention that so many fear.
AI Isn’t Dangerous. Human Error Is.
Let’s start by talking about the most common depiction of AI in popular culture. The rogue system that takes its original purpose to the extreme, either putting human lives at risk or creating an outright doomsday scenario. Ultron is probably the best example in recent memory, but there are plenty of others.
Skynet. Hal 9000. AM. Cylons. The Borg. The Matrix.
You get the idea. In actuality, this sort of scenario is extremely unlikely, if not outright impossible. In a recent interview with UC Berkeley, Ph.D. computer science candidate and machine learning specialist Dylan Hadfield-Menell shed a bit of light on the matter.
“The innate qualities of AI are largely misunderstood by the general public, leading to a distorted understanding of the risks,” Hadfield-Menell told author Glen Martin. “Rogue AI often … achieves sudden awareness and then decides to kill all humans. That’s not the real problem.”
“It’s more a matter of motivating AI improperly or the AI employing unexpected means to achieve its programmed ends,” he continued. “Take a [computer enhanced] vacuum cleaner that you’ve programmed to suck up as much dirt as possible. It quickly sucks up all the available dirt, so at that point, it dumps out the dirt it has already collected so it can suck it back up again.”
In other words, we’re unlikely to ever see actively malicious AI. Instead, any AI-based catastrophes will ultimately be rooted in simple human error. That doesn’t mean we couldn’t eventually see apocalyptic AI.
Just that if we do, it’ll be the result of a series of human blunders rather than sudden sentience.
AI Isn’t Going To Replace Human Intelligence
Another common fear around AI is that machines will ultimately and inevitably replace people. Again, this is unfounded and based almost entirely on misinformation. The reality is that human intelligence and machine intelligence are wholly distinct, with each complementing the other.
Even if we do somehow manage to program sentience and self-awareness into our AI algorithms, they will still see the world differently from us. We will likely still be more formidable than any machine in areas of creativity, emotional intelligence, and long-term planning. In even the most advanced supercomputers, machine learning functions differently from human learning.
“It would be a mistake to say that [AI] algorithms recreate human intelligence,” David Watson of the Oxford Institute and Alan Turing Institute recently told Forbes Magazine. “Instead, they introduce some new mode of interference that outperforms us in some ways and falls short in others.”
Will AI-based automation make certain careers obsolete? Yes. Does that mean AI represents some sort of looming job crisis? Hardly.
AI Is the Holy Grail of Analytics
Let’s wrap up by looking at the other end of the spectrum of misconceptions. The notion that through machine learning, a business can gain an unchallenged competitive advantage over its peers. AI-enabled analytics can offer some incredible insights, true.
But you have to look for those insights. You have to know what data to feed into your algorithms and why. You need to give your analytics platform direction and guidance.
Otherwise, you’re simply crunching gigabytes, terabytes, or even petabytes of meaningless data.
While it’s still incredibly powerful and fascinating, AI is neither a looming apocalypse nor is it a solution to all your business’s problems. The reality is far more nuanced. At the end of the day, the advent of AI really won’t be that different from the invention of the Internet or the birth of the smartphone.
Industries will be disrupted. Countless new innovations will hit the market. And the world will keep marching forward, just as it’s always done.
About the Author
Max Emelianov started HostForWeb in 2001. In his role as HostForWeb’s CEO, he focuses on teamwork and providing the best support for his customers while delivering cutting-edge web hosting services.