Movies like Blade Runner and Her have popularised the idea of fully conscious computers, and with AI (Artificial Intelligence) technology like Apple's Siri or Amazon's Alexa increasingly present in our lives, it'd be easy to believe that what you see on the silver screen is just around the corner.
Whilst I enjoy a Sci-Fi epic as much as the next person, in my dual role as Professor of Computer Science at the University of San Francisco and Chief Scientist at data integration software provider SnapLogic, I investigate the practical applications of AI and am tasked with explaining and teaching the realities of what can be achieved. In other words, I separate the fact from the fiction, which is what I aim to do today.
A Conscious Computer?
While Alexa seems close to a computer that can interact like a person, the reality is that Alexa (and any computer program to date) is not what we'd call conscious. It's not self-aware or able to generate original thoughts.
What many people call AI is actually a subfield called machine learning (ML). It's grounded in statistics realised in computer programs and combined with the data of previous observations. At times, these smart devices and programs can seem intelligent, but are simply a combination of code and past data which provide a prediction or recommendation.
Take a self-driving car. Autonomous cars attempt to follow the rules of the road and replicate a human's driving style. From the outside that can appear like it's doing something in the same way a human does, but it isn't conscious. The computer code in the car is carrying out a directive (drive to Starbucks) by creating a plan based on digital maps and then responding to the surrounding environment to navigate to the destination.
Augmented Intelligence and Rise of the Robots
AI akin to a fully conscious person is still the stuff of fiction, and not the goal of many computer scientists in the field. For many the future of AI is really the augmentation of human intelligence in very specific situations.
In this future world, computers will assist humans in their daily activities and their work-related tasks, and this is already happening. For example, Google Maps dynamically suggests alternate routes for you.
But will we allow machines to make decisions for us? In some situations, yes. Smart cars can now determine if the brakes should be applied to avoid colliding with a vehicle that has rapidly slowed down. However, as the designers of this functionality we control how and when it works. It's possible that the automatic braking system may brake unnecessarily, but not because the computer chose to do something different, but because the algorithms responded incorrectly to the incoming sensor data.
The main point is that all of the code that appears as intelligent is ultimately under human control. A computer program that is self-aware and chooses to make decisions that deviate from the intended behaviour, such as in Ex Machina, is unlikely to happen any time soon.
That said, if I'm wrong and if computer scientists can develop programs that achieve truly conscious thought, it's not clear to me that such an AI would naturally present a risk to humans. If possible, it seems to me that a self-aware form of AI would have the same incentives to cooperate with other forms of conscious beings, i.e., humans. The fear of AI robots taking over the world, while I suppose a possibility, seems to be one possibility of many that don't involve total robot domination.
Will Robots Take My Job?
In certain fields such as manufacturing, AI may replace human workers or parts of their job, but the idea of AI eliminating entire job categories? That's not likely any time soon.
Rather than simply replacing our jobs wholesale, AI is going to shift some of them dramatically. All those boring tasks you hate doing in the office? Let AI deal with them. AI will deliver them more quickly for businesses, allowing them to channel resources towards other areas which require more human skills - emotional reasoning, complex manipulation, and creative intelligence.
AI is currently supporting humans in their tasks, and slowly but surely, as the technology develops, its influence will spread to allow us to focus on only those tasks we choose, namely those we're best suited for and where we can add unique value.
Remember how, when researching a topic, we had to venture to the library and trawl through books for information? Now we have Google, which does the job more quickly for us, allowing us to focus on the task at hand. That is AI and machine learning at work.
Man and Machine
At the end of the day, while computer scientists are pushing the state of the art in AI and ML, we still have very little understanding of how human intelligence really works. As scientists continue to delve into this, further applications of current AI and ML techniques will continue to augment the many facets of human activity. But, it's not going to be able to replace human activity, at least not for a very long time.
For now, the robo-future remains rooted in fiction, not fact.