The world dreams of a free market where every individual can contribute to the society and economy in his or her own capacity. But the downside to every beautiful thought is the presence of the ugly truth. The world economy is monopolised and free-trade is almost non-existent with big conglomerates and corporations calling the shots.

And this trend has started to trickle into the technology and innovations sector as well. Blue-chip firms and Big Data giants have started to monopolise the market because of the resources and data at their disposal. Not only are they controlling the markets, but have also managed to develop technology that surpasses its competitors by miles!

One such software that is on the radar yet relatively unspoken about is the AlphaGo AI that was created by the Deep Learning Company ‘DeepMind’, which was acquired by the data giant Alphabet (The Company that owns Google).

You might be wondering about the big deal behind a company developing an AI. Every other company or tech-startup is creating one.

It’s the magnitude of effectiveness that has sparked off the debate amongst closed circles. The AI, that was created by DeepMind, was named ‘AlphaGo’ and was designed for a single purpose only and that was to play the ancient Chinese game of ‘Go’, and master it.

Now before we get into the details further, we must understand the complexity of the game. The game is played on a board and is comprised of two coloured stones, black and white, played along latitudinal and longitudinal lines by two players. The objective of the game is to capture the opponent’s stones. The permutations and combinations that lead to the end of a game is more than the number of stars in the sky. Those are the odds of the game!

The initial AI was called ‘AlphaGo’ and it consisted of ‘Black and White Pieces’ data as well as a few hand-engineered techniques and strokes to play the game. From this data, the neural network self-taught itself the game of Go and became a formidable player, defeating the 14 time world champion 100 games to 0. The second stage of the AI was called ‘AlphaGo-Zero’ and it consisted of a variable neural network that predicted the right moves, instead of playing a 1000 games in a trial/error fashion like its predecessor AlphaGo.

AlphaGo-Zero became so good at the game that it surpassed AlphaGo to become the best player in the world and defeated the best players by devastating margins.

This was a success story! The AI AlphaGo-Zero became a self-sustainable and ‘intelligent’ entity in the game of ‘Go. One of the highlights of the game was the unprecedented ‘Move 37’ when the AI was playing against world no.1 Lee Sedol from South Korea. When the move was made, Lee exclaimed “That is a very strange move” and the move went on to change the course of the game as the AI won within the next few moves. The AI predicted the future of the game though ‘Move 37’.

But we haven’t reached the scary part as of yet. AlphaGo-Zero became so smart and self-aware that in a matter of time, it was able to develop into an all new stage, and the unthinkable happened!

AlphaGo-Zero created its own AI that could play ‘Go’ as efficiently as itself. The software was able to isolate the successful source code and duplicate it. It basically created its own AI which was as smart as itself, if not smarter!

The ability to create its own AI perplexed the Google team and hence, is kept under wraps. The AI was only fed with data on how to play ‘Go’. The potential complications that could arise by feeding the AI more data, such as defense access or even military data has urged Google to put the release of its AI on hold.

Elon Musk, CEO and Founder of Tesla and SpaceX, has urged the world’s top companies to control their AI roll-outs until further research is carried out and the technology is understood better, so as to prevent the creation of something ‘unsafe’, like the AI AlphaGo-Zero. But the vehement opposition due to a demand for a return on investment (ROI) from investors has led to a polarization between the parties.

“There’s a lot of risk in the concentration of power. So if AGI [artificial general intelligence] represents an extreme level of power, should that be controlled by a few people at Google with no oversight?” – Elon Musk.

So to democratize the process of Artificial Intelligence construction, Elon Musk and a group of supporters launched an online program known as ‘OpenAI’. This project is supported by Infosys and Microsoft as well, and is a non-profit project for the generation and creation of ‘safe AI’.

This revolution to create a free market for AI development comes in the wake of the fear of a certain unknown. But forums and conferences are being created in order to generate awareness and educate the people on the advancements as well as safety precautions needed to be carried out while creating an AGI. One such event is the World AI Show, slated to be held at The Address Dubai Mall on 11-12 April 2018.

The event is being conducted by Trescon. For any details on how to attend the event, you can visit the event page World AI Show or directly contact:

Nikhil Menon
[email protected]

SOURCES:

• Paulose Kocherry Jolly Thompson – Small Business owner and AI enthusiast

• http://www.independent.co.uk/life-style/gadgets-and-tech/news/elon-musk-artificial-intelligence-openai-neuralink-ai-warning-a8074821.html

• https://deepmind.com/blog/alphago-zero-learning-scratch/