From Hal in “2001: A Space Odyssey” to “Westworld,” the thought of emergent artificial intelligence chills us to the bone. But is AI technology really a threat? Or does that only happen in science fiction?
Some, including titans of the tech industry like Elon Musk, are sounding the alarms. They’re even calling for (gasp!) the U.S. government to put more regulations in place. That’s startling, coming from a devout, self-proclaimed libertarian. Meanwhile, some experts in AI technology dismiss the emergent artificial intelligence threat as the stuff of science fiction. Nuerala CEO Max Versace even lashed out at Musk for “selling fear.” Then again, he just happens to run an AI and robotics firm.
Artificial Intelligence vs. Emergent Artificial Intelligence.
First, let’s get our terms straight. We have the everyday robots and “bots” who help us with tasks in our daily lives. And then we’ve got the creepy would-be robot overlords from science fiction.
Artificial intelligence is the theory that we can build computer systems that mimic human (natural) intelligence. Key areas include learning, planning, reasoning, problem-solving, language, and perception. There’s also being able to move and manipulate objects, and what’s called “knowledge representation.” In a nutshell, that term means the ability to know things, including when you don’t know something.
AI tech also seeks to develop “social intelligence” so bots can read people better. They’re also working on creativity for going beyond what a system’s programmed for. This might make voicemail trees more helpful and less annoying, but it’s also creepy as heck. No wonder some of us are worried.
Still, AI experts assure us true, sci-fi like emergent artificial intelligence would require a sense of self-awareness. Without that, even the most human-looking and socially adept cyborg can’t change how they’re programmed. In other words, we’re a long way from Westworld.
— Unity motion leap (@EvertRob) December 1, 2017
Coming up on the 50 year anniversery of 2001; A Space Odyssey. One of the most prophetic important films in cinema history. We are warned about AI. https://t.co/gYHKHuJA0g
— satex (@satex) December 1, 2017
But where do you draw the line?
Some disagree on where to draw the line for emergent artificial intelligence. After all, where does the process of becoming aware of yourself begin? Most of us see it as the moment when a robot knows it’s a robot. Yet some argue the self-stacking blocks below are self-aware. Why? Because they know how to find each other and cluster together in a straight line.
— James Romero (@MrJamesRomero) June 17, 2016
Most of us are okay with having Roomba clean our floors or asking Siri for directions. What freaks us out is the looming specter of emergent intelligence … The idea that robots could become free agents with thoughts, ideas, and desires of their own.
Examples of AI Tech in Use.
In early 2017, Forbes listed the most common uses for AI tech. They also note that as AI comes into its own, it remains in a fairly early stage. And anyone who’s ever cursed at their autocorrect knows, we’re a long way from the kind of emergent AI people fear.
Here are the eight most common uses of AI Tech today.
1. Personal assistants: How do Siri, Alexa, Cortana, and other voice-activated apps start knowing what you want? Through a process called machine learning. Machine learning lets these apps learn things they weren’t specifically programmed to do. That’s how they learn to do things like give us directions, add things to our calendars, send messages, make phone calls, and find information for us.
2. Autocorrect: The autocorrect apps we use on our smartphones are programmed with a dictionary. They also use machine learning to figure out how we — as individuals — type our messages. There are even apps that transcribe our voicemail messages and send them via text. These apps are growing more and more accurate.
3. Smart cars: Today’s smart cars help us get better fuel economy. Plus, they give us directions and traffic info via GPS. Oh, and they also connect to BlueTooth and other wireless devices. Google’s Waymo and GM even have some self-driving cars on the road. But Engadget reports the smart cars of the future will do more. New features will include “self-healing paint” to make those scrapes and dings vanish. Augmented reality “head up” displays on your front window are also in the works. That way, you can access directions, your speed, and incoming calls … without ever taking your eyes off the road.
4. Customer service: Software like Cogito uses something called “behavioral adaptation” to improve their customer support. These bots listen in on calls and learn to detect when an interaction’s about to go south. They can then alert the customer service rep so they can take a different approach. Other bots like eGain’s AI Chatbot work with customers directly. They reduce wait times by helping with basic issues. That way, live agents only need to handle the more complex transactions.
Can AI really improve customer service?
— eGain (@eGain) June 30, 2017
5. User experience: Boxever and John Paul are innovators in the travel industry. Their software seeks to provide a high level of personalization to excite people and keep them highly engaged. The video below shows Boxever gets people who’ve bailed on booking their trip to complete their transaction. Is it helpful or creepy? Here’s the video so you can decide.
— Yves Frateur (@YvesFrateur) June 4, 2015
6. Transactional AI: This is the AI tech behind Amazon’s and Netflix’s uncanny ability to know what you’ll like. They gather data on what you search, browse, buy and use it to make recommendations. Pandora radio takes this to a whole new level with “musical DNA” that keeps them serving up music users love.
7. Home tech: Smart home products make it easy to control your lighting, thermostat, home security system, TV, and various appliances with your smartphone. Amazon Echo and Google Home even function as a control center for your music and TV.
8. Games: Games have long mimicked emergent artificial intelligence. For example, Portal’s operating system and main character, GladOS, starts out as a guide. But as you level up, she grows more and more malicious, giving bad and often lethal advice to the player.
Free AI experiments you can try for yourself.
Now, game developers are starting to get serious and Google’s created a playground for them … And for us. Their AI Experiments Program features various fun programs that let us teach them and also lets those of us who program submit our own experiments. So far, these AI technologies include a music program that can harmonize with us when we sing, a sketch program that transforms our rough drawings into simple but professional-looking renderings, and more.
Is emergent artificial intelligence a threat?
If you’ve used any of the above AI technologies, you already know AI tech’s got a long way to go before it’s a threat. Tesla’s Elon Musk and UK theoretical physicist Stephen Hawking keep warning us of the dangers. But Microsoft’s Bill Gates and Facebook’s Mark Zuckerberg said it’s too soon for us to even know what we should regulate.
But Jeremy Straub, an assistant professor of Computer Science at North Dakota State University, argues that we’ve already got plenty of laws on the books that govern AI. First of all, drones have to obey FAA regulations, and self-driving cars have to obey traffic laws.
Yet Carlos E. Perez, a software architect and author of the “Deep Learning AI Playbook,” and a software architect makes an interesting point. He writes on Medium that it’s never too soon to start pushing for some reasonable oversight. Given the glacial pace of the U.S. government and the current anti-science administration, better sooner than later.
Perez cites concerns about corporate monopolies. He also warns of “weaponized AI in the domain of disinformation and cyber-warfare.” For example, the possibility that Facebook allowed trolls and bots to influence the 2016 election should raise alarms. Also, if AI won’t destroy us anytime soon, it’s definitely killing jobs.
Perez refers to Rachel Thomas from Fast.AI, who decried “Musk’s fixation on evil super-intelligent AGI killer robots in a very distant future.” But to her, the mass unemployment caused by AI replacing jobs poses a far worse threat. That, plus what she sees as too much wealth flowing to the top could cause major social unrest.
How do we feel about AI? You’ll Be Surprised
Meanwhile, PEGA Systems — which makes adaptive software that streamlines sales, marketing, and customer service functions for businesses via the cloud — took a survey on our hopes and fears when it comes to AI tech. Respondents were divided about letting businesses engage with them using A third said they’re “comfortable” with that, another third said they felt “uncomfortable” with it. And the remainder said they’re “not sure.”
Yet the study also found people were more accepting of AI when they realized they’re already using it. It turns out 84 percent were using AI services and devices — like Alexa and Google Home — but didn’t know it. Sure, 74 percent said AI makes them feel uneasy, and 24 percent “worry robots will take over the world.” Yet 68 percent still said they’ll accept it if it makes their lives easier.