Can we meet HAL 9000 in our life times

In pursuit of AGI – The strong AI

Human race is evolving with every passing day and what’s evolving alongside is the usage of AI. From Alexa becoming the new member of our family to Driver-less cars to OTT Platform recommendations, we are well surrounded by algorithms. AI has become the household name in last decade and has advanced by leaps and bounds. These advancements give lot of hope to the old classical dream of creating human intelligence using algorithms. While we have been focussing on the Weak AI there is another side of AI that is referred by the researchers as The strong AI which is called by different names varying from Strong AI to AGI to Full AI.

Roots of Strong AI goes back to the fictional character ‘HAL 9000’ in the movie ‘2001’ where Arthur C. Clarke, the Antagonist, has showcased the glimpse of possibility that could open if we build strong AI. However the real advent of strong AI was when AI pioneer Herbert A. Simon wrote in 1965: "machines will be capable, within twenty years, of doing any work a man can do."

Even after five decades, we are still struggling with the same questions around Strong AI. There are some strides which we have made in this field yet we don’t have major breakthrough. Many pioneer educational institutions are collaborating and researching in this space - MIT presented a course in AGI in 2018 - but still the light in the tunnel seems too far.

There are two sets of views which we get from the research community. On one side we have Henry Markram, director of the Blue Brain Project, who says “It is not impossible to build a human brain and we can do it in 10 years,” on the other side we have Rodney Brooks, an MIT roboticist and co-founder of iRobot who says that AGI wouldn’t arrive even till 2300. While we can sense the contrast of optimism and pessimism in the above views, the one which sounds more realistic is the statement by Richard Sutton, professor of computer science of the University of Alberta who says “Understanding human-level AI will be a profound scientific achievement (and economic boon) and may well happen by 2030 (25% chance), or by 2040 (50% chance)—or never (10% chance).”

The majority of researches which happened in this space are focused on solving the following

Human brain Processing: Each of the 1011 (one hundred billion) neurons has on average 7,000 synaptic connections (synapses) to other neurons. A three year old child has nearly 1015 synapses which is equivalent to 1 quadrillion. Though this no is lower in grown-ups but still it is enough to require ultra-super computer to process that. The processing power of brain is assumed to be 100 trillion synapses updates per second as per few estimates it is believed that to equal the human behaviour we need to have the computer which can do 1016 computations per second which is equivalent to 10 petaFLOPS. The good news is that we have got the world’s most powerful supercomputer created in Japan which could reach the 415.53 petaFLOPS on the LINPACK benchmarks. So this limitations seems resolved but then we have few more to add to the list.

Consciousness: Even if we could address the synapses updates per second, there is a larger discussion on the generation of subjective experience and thoughts. As we discuss more on this the topic, it leads us to the philosophical route but it would be interesting to see how Strong AI will include the subtle aspects like self-awareness, sentience, and sapience

Sentience: Although we have made lot of advancement in the space of deep learning, the dream of creating the replica of human brain still seems farfetched. Whether we consider the Computer Vision or Neural Networks, both are evolving presently and will take their own time before they mature

The list of challenges is not exhaustive, in the process we have overcome few of these and researches are going on to come up with the solution for others. The overall progress so far in the areas of deep learning, computer vision and the likes has rekindled the hopes of scientists worldwide but tackling emotions, consciousness and awareness aspect of it still remains a challenge.

As we move closer towards the breakthrough in the Strong AI, businesses should start focusing on the researches happening in this area and should collaborate with the researchers, academicians. The benefits of such collaborations would be mutual. While businesses could fast pace the research programs by investing into them, the breakthrough IPs created in this journey can help businesses in building products and keep pace with the evolving technology landscape