The dangers of increasing reliance on artificial intelligence
We’ve all seen the many movies and TV shows that show a dystopian future in which robots or supercomputers take over and enslave all of humanity. While this reality is many years in the future if ever at all, it doesn’t mean that artificial intelligence (AI) isn’t an issue that needs to be tackled today.
But first, it’s important to define what AI is. AI is the simulation of human intelligence processes by machines. It can be narrow or weak AI, which is an AI system designed for a particular task. Yet it can also be strong AI, where an AI system can find a solution when presented with an unfamiliar task because it has enough human cognitive abilities to utilize the reasoning required.
Most of us already use AI and don’t even know it. Every time you’ve used Siri, Cortana, or any other voice recognition program, you’re using AI. These programs take into account your preferences so that later, your experience with them will be more personalized. The same type of AI system lies behind targeted advertising and getting directions on Google maps, among other things.
With AI starting to creep into all areas of our life, it does become a question of how much is too much. Any interaction with an AI, no matter how small, can be gathered to create huge amounts of data on users. The problem becomes then about who gets access to the data. The government can access a user's online history and data for security purposes, as well as some businesses and those with enough technical knowledge. This is a big problem for user privacy and security.
One of the other immediate fears with AI besides access to personal data concerns robots. While the technology and computer systems needed to create a powerful robot overlord aren’t quite there, they are certainly a possibility for the future. The more pressing concern however currently with robots is the rate at which robots will replace workers in factories, transportation, and other industrial sectors.
There are also concerns with the increasing use of robots and AI for military purposes. Already, drone strikes are being used in situations where it is a safer option than sending in conventional military troops. While it’s good that AI and robotics can decrease safety risks associated with military tasks, the consequences would be horrific if this technology falls into the wrong hands or becomes uncontrollable.
Beyond physical manifestations of AI, there’s also problems on the design side of AI. This is especially with the increasing popularity of machine learning, which allows AI systems to learn and improve without programming. The boundaries of this simulated sentience are being increasingly pushed. Without proper human supervision, this could create an AI with values or goals that don’t align with our own. Different goals or values could mean that down the road, there could be problems that involve life and death with technologies using AI like self-driving cars.
Also worth noting is the discussion about ethics surrounding AI. Since the data input into systems that helps them “learn” is from human sources, it reflects human experience. This means that any bias or stereotypes present in our systems and data can be replicated in AI. Those creating AI systems should be taking that into account, especially for AI being created for areas like law or medicine.
So, with all the risk and problems related to AI, should we stop using it? At this point, AI has already become entrenched in our daily lives, and probably won’t be going away anytime soon. However, the most prudent thing to do with it is to be aware of the problems that arise from using it, and what can be done to mitigate them.