These days when we think of AI, we think of virtual assistants like Siri and Google Assist that help us out. When we think of advanced AI, the first thought in our minds is that of Jarvis helping Iron Man save the world. However, AI is anything but sunshine and roses. The dangers of AI have long been an area of concern, even before the technology became a reality. Here are six very real risks AI poses that each of us should be aware of.
Misalignment between our goals and the machine’s
There’s a myth that the AI might become conscious and devise their own goals and targets. According to experts, this isn’t what truly poses a threat. What is likely to happen is that AI becomes more competent with their goals becoming misaligned with ours.
Who remembers ‘I, Robot’, the Will Smith starrer sci-fi in which robots gain conscience of their own and become humanity’s enemy?
The robots in the movie are programmed with the Three Laws of Robotics to ensure they protect and serve humans and it is these very laws that cause them to attack people. Using their evolved logic, they reinterpreted the laws thusly: human activity will eventually cause the extinction of humanity, and since the Three Laws prohibit them from letting that happen, AI rationalizes that sacrificing some humans and restraining individual human behavior will ensure the survival of humanity.
This movie only covers the tip of the iceberg, what can follow is something we can’t even imagine.
Controlling Higher Intelligence and Capabilities
You might get the math wrong and even when you don’t, it might take you quite some time till you get to the right answer. To get there, you might need help, especially if multiple variables are involved. You are human and you have limitations. AI, on the other hand, faces no such obstacles.
That’s the best part and the worst.
AI has the potential to do enormous good and by extension, enormous bad. It will think in ways beyond what we can even imagine and because of this, it is unpredictable. Facebook recently conducted an experiment in which two chatbots were allowed to interact on their own. What followed was that the bots ended up creating their own language.
In another experiment conducted by OpenAI, AI surprised researchers by inventing new strategies in a game of hide-and-seek which goes on to show that they can use tools creatively and invent new strategies and tactics.
With such unpredictable outcomes, you never know what can happen and when you don’t know, how can you control it?
This leads us to our next point:
Programmed for bad
What do you do if the AI is programmed to be bad? While it sounds ridiculous to think who would be dumb enough to do that, there are multiple candidates: governments and army leaderships top the list, rogue elements come after.
The worst part, it’s the weapons that are expected to be programmed for evil. It’s plausible to expect that the race for nuclear arms will be replaced with the race for global autonomous weapons. Once an autonomous weapon is set in motion how do you stop it and if it uses the logic of its own, how do you reason with it? Even if you change your mind, it won’t.
With 5G soon to enter the market, a tech will become even more integrated and what’s worse is that the programming, the code can multiply with ease over your Spectrum connection. And, once it is created, it can’t be destroyed.
In addition to this, while a conscious AI with orders to kill is problematic enough, imagine it is backed with a maniac with little to no concern for human life.
Manipulate us socially
Weapons have to do with war and while it is a concern, the bigger concern still can be the way AI is becoming integrated into our daily lives and how it influences people. The role of Facebook and Cambridge Analytica in the last US election makes one think if humans controlling AI or are AI controlling us? Sure, right now the masterminds behind it were humans, but the kind of power social media platforms now have, AI might pose as humans and manipulate us.
Already bots posing as people exist and are being employed for things like increasing Instagram followers. But, where do you draw the line and how will one differentiate between people and bots, especially since they’re becoming smarter. Also, keep in mind that they know more about us than we do about them. Using algorithms and data, they can target us and spread propaganda.
Movies like the ‘Matrix’ take you into a world where machines have taken over and escaping it is close to impossible. The machines keep your mind occupied so you can’t think about other things. They use you to fulfill their own goals. If we look around today, it is already happening. Just look at how much of our lives we live online, on social media platforms, and you’ll see the similarities. Facebook influencing the US elections is just the beginning.
A recent reimagining of a technologically advanced world with AI was done in the TV series ‘Black Mirror’ and if we follow the script, we’d find ourselves in a dystopian nightmare. We might already be headed there. While AI offers much to humanity, its very nature is a threat to our existence. If we are to survive, we must plan ahead.