Hold on.. The Next World is here....

Artificial Intelligence: A reminder for us to swim in deep water

Mark Zuckerberg during the Facebook F8 Developers Conference

Artificial Intelligence is not scary as it appear in popular culture. We have control over it and humans are meant to lead, not to work.


AI is dominating one thing at a time.
Deep Blue, a chess-playing computer developed by IBM, became the first computer to win a chess game against a reigning world champion (Garry Kasparov) under regular time controls back in 1996. In 2016 Lee Sedol, winner of 18 Chinese-Go world titles and widely considered to be one of the greatest Go players, lost 1-4 by Google Deep Mind’s project, Alpha Go artificial intelligence.

When these events took place, news providers did their best to instil the feeling of doom into our minds. At the time CNN news read, “Google’s AI just beat the world’s best Go player”. Just to make things worse, Alpha Zero – a newer version of Alpha Go – managed to beat Alpha Go without the need of humans (Alpha Go required human intervention to train on the game)

image of Lee Sedo
Lee Sedol. Image Source

“It’s scary how quickly it upgraded, it feel like its, Terminator” said Korean Go-Master Kwon Kap-yong in an interview.

Machine Learning takes time and effort
Machines do not learn by itself like humans do. Less sophisticated method in machine learning is supervised learning. In supervised learning the software requires human expertise to optimize the parameters and a fair amount of data to practice itself on. The more sophisticated methods such as neural networks do not need much guidance as supervised learning, however it requires a lot of computing power and training data to compensate for that missing element of human guidance.

And one important fact to point out is, humans do learn and take actions simultaneously whereas for AI, it is a two-step process. Therefore artificial intelligence cannot be deployed in tricky uncertain situations. Which renders human characteristics invaluable in situations where leadership is required.
Even Alpha gets beaten

In Lee Sedol vs. Alpha Go match, Lee actually won one round out of five. After the match Lee mentioned he identified a weakness in the AI’s strategy and leveraged it to turn the game in his favor. He even mentioned that he could have won two more games earlier if he capitalized on Alpha’s errors properly.

Where is the weakness?
Remember computer chess games that we used to play against the computer back then? The program difficulty level is determined by the extent of possible moves evaluated by the opponent and the program. The problem is number of possible moves increase significantly after the first move. After the forth move there are 288 billion possible moves. It is impossible and wasteful trying to evaluate every single move against us.

That’s where the AI’s weakness comes in. Super computers are faster in processing data however not intelligent enough to select what to process for the decision making. Effectiveness of decision making relies on what data to process.

This is a tricky problem as an AI can be over-optimized where the AI will simply wait for the best conditions to make the decision which might occur rarely or never. Or the AI can be under-optimized where the system will make more inaccurate decisions.

AI’s speed is compensated by our intuition.
Striking the perfect balance between over and under optimization can only be achieved by our own intuition so far. AI has rather an arbitrary methods on selection whereas animals/ humans are equipped with few hundred million years’ worth of intuition-sharpening took place during evolution.

An Image of baby cheetah
mammals are born equipped with knowledge of “intuitive physics.” Image Source

The formula of Success
The basic requirements of success as defined by countless number of successful personalities commonly include courage, perseverance, ingenuity, etc. These qualities revolve around to beat odds in an unpredictable environment. Carefully calculated and over-specified plans fail most of the time due to poor adaptability in ever changing circumstances. Intuition is required to circumvent these problems something that machines clearly do not have.

We are the gods of our own creations: The kill-switch
No matter how much perfect the automation will be, humans will never trust a machine completely unless we completely understand how the mechanics behind an artificial intelligence works. You can expect the person who created the AI to let it take full control over his life. Remember, many inventors thought they fully understood their inventions, only until the moment they were killed by their inventions.

Take autopilot in Tesla or radar mode in Mercedes as an example. They are perfected to control the vehicle in highways with more caution and vigilance than a human being. However the human driver never really know what is processing in their digital minds. That’s the worry made car manufacturers to put a kill switch by disengaging the AI when the driver turns the steering wheel.

Humans want a kill-switch for one reason more than anything else. It’s the need of ultimate control over everything. We are the gods of our creations, the creations cannot take over their gods.

European Parliament’s legal affairs committee voted to pass a report urging the drafting of a set of regulations to govern the use and creation of robots and AI. Image Source

We are looking at it the wrong way
It should be obvious now that and AI cannot assume a leadership role in real world. It means humans can take on being leaders and key decision makers while letting the robots do the middle-management and carry boxes around.

AI is finally paving the way for humans to exclude themselves from mediocre tasks and to reach their maximum potential. If billions of people work on their ingenuity while robots do the deliveries, cleaning and managing mediocre tasks, the human kind would advance as a civilization at an unprecedented speed ever witnessed by history.

Moving to deep waters
Flooding scary and unsettling news about Artificial Intelligence has revealed how uneasy the general population feels about not being content with their lives and reach higher levels. The AI has already pushed humans swimming at the shallow end to the mid point few years back. In few decades the machines will inevitably take over the middle ground as well. The next generation of humans will be forced to live their lives reaching the maximum potential swimming in deep water. That future is far too ahead to predict how the humans will feel about that.

Leave a Reply

Your email address will not be published. Required fields are marked *