When I had first read about Artificial Intelligence, it had seemed like an improbable science fiction concept. Robots thinking, learning, feeling, and taking decisions on their own still seems a little out there.
But, don’t be fooled – robot or not, you are surrounded by AI today in your everyday lives.
No matter which smartphone you are using, you are most likely to have interfaced with a virtual assistant – the likes of Siri, Allo and Cortana. Per Microsoft, Cortana “continually learns about its user” and would, in time, anticipate the needs of the user. We are in an age were cars are driving themselves, news is being generated by computers, banks are sending automated fraud alerts – in some form or the other, artificial intelligence is everywhere.
Despite all the different ways in which AI is making our lives better, many tech giants are concerned about the ill-effects of the technology! Stephen Hawking, Elon Musk and Bill Gates have all openly expressed their worries about super-intelligence.
In a Reddit AMA, Gates clearly mentioned, “I don’t understand why some people are not concerned”.
In 2016, Google published a technical paper on the key concerns with AI, titled Concrete Problems in AI Safety. Here we present you a brief synopsis of the same, with sinister conjectures on how the development of full artificial intelligence can wipe out humanity:
Kill to Improve?
Let’s imagine an AI that is continually learning to do one particular action. Over time, it will start experimenting with ways to do its duty, to improve efficiency. But, what happens when the best way to perform its task adversely impacts something else in its environment? The Google paper gives the example of a cleaning robot – will it knock over a vase because it can clean faster by doing so? But, as you might imagine, the example can be much grimmer.
Would an AI value a life over doing its assigned task more efficiently? Will it kill to be better? Horror show, right!
Also, during such explorations, the system might blow up its environment, purely accidentally. So, it is kind of like a busy toddler without adult supervision. All the parents in the audience are going googly-eyed right about now, I bet.
Taken over by the Evil Forces
Ok, so we are letting AI into our lives, letting it control us, for convenience. Our cars are taking us to places unsupervised, our kitchen equipment is learning our food preference and serving independently, our home is keeping itself secure – all because of crafty codes. Now imagine being hacked.
In the future hacked might not just mean embarrassing photos sent over your email or all your money spent on a hot tub in Thailand – it might mean the difference between your life and death.
God forbid, if ‘evil forces’ can run your car, or your home security, that is not a warm, fuzzy feeling – is it?
An AI is smart – right! And, it can keep getting smarter – unaffected by needs (or stupidity!) that keep humans occupied. Imagine an AI that has the simple goal of getting more intelligent, and is not thwarted in its quest by the ‘Cat poops, Baby sings’ video. What will then stop it from overriding human intelligence by a mile and making us obsolete?
It is not a debate anymore that super-intelligence would emerge, although the date remains unpredictable. Some say it is impossible before 2100, others that it will happen by 2060.
To be honest, the worry is not that AI would become evil or conscious. The real worry is whether the goal of AI would be aligned to human goals, or divergent from ours. In the latter case, we are in big trouble! If you want to get scared some more, watch Black Mirror on Netflix, you won’t chill, I promise.
I find immense relief in the fat that the biggest technical minds are on it. Here’s hoping they will find a way to keep humans from becoming puppets to steel creatures.
What do you think of AI? Comment below or write to us at firstname.lastname@example.org
- Earth Day 2018: Say NO to Single-Use Plastic Products, NOW! - April 21, 2018
- #EqualPayDay – 4 Conversations to close the Pay Gap - April 9, 2018
- World Health Day: Can India achieve Universal Health Coverage? - April 7, 2018
- India’s Caste Malice: Our Failed Crusade against Apartheid - April 3, 2018
- Bangalore’s Water Crisis: The Economy of Thirst - March 5, 2018
- Lathmar Holi and India’s problematic relationship with Stalking - March 2, 2018
- Are Books and Bookstores Dead: What we learned from the West! - February 23, 2018
- Beyond Chapati: Exotic Bread and Bakes of India - February 14, 2018
- Why you can’t police Valentine’s Day out of India - February 14, 2018
- WhatsApp India: Why Indians are falling prey to the Fake News Mafia - February 7, 2018