AI and its promises

Recent marches in technology with great focus on cloud computing has brought AI to the market forefront. Every organisation appears eager to portray itself as AI ready and AI friendly. Buzz words are thrown around as a gimmick to attract potential customer. Such was my experience from a recent conference titled Customer For Life” organised by The Economic Time.

The conversations around AI seem to be full of positivity and energy. Countless benefits of a godlike benign, thinking machine are recounted at every opportunity to sell. The problem with the current discourse is that most people talking about AI are managers and salespersons who know little about the underlying technology. With the sellable positives getting the limelight and start-status, the unresolved questions of ethics and social change are left in the backburner- delegated to people in academia to solve.

The situation and mass belief system of worshiping AI can be described as “irrational exuberance” as used by Bob Schiller in reference to the subprime crisis of 2008. The over enthusiasm of acolytes deifies questionable and unethical practices. While Schiller expanded on the term in a narrower financial domain, the conduct of human mind in groups has been shown to exhibit irrationality in almost every bubble the world has seen. The questions that need answering today in relation to AI, however, are not ones that we can let organisations or governments decide. The best case scenario would be a public debate in a world where every citizen is a subject expert. This world, however, comprises people from all walks of life. And to complicate matters exponentially- this world is in the middle of a full scale crisis. Overpopulation has led to resource crunch and massive stress on the environment. People today are desperately looking for immediate solution and will follow any cult (religious or scientific) to dispel such fears for borrowed peace of mind.

Enter AI. An all-powerful mind that can outsmart all of smartest people that ever walked the earth-Combined. Ray Kurzweil in his book titled “The singularity is Near” talked about current times been a knee of a exponential growth curve of intelligence. He argues that while our past has seen a linear growth in technology, once silicon intelligence gains the critical mass, the world will see levels of growth unheard or unimagined in the past. While projections of the future are always riddled with uncertainty, the past 20 years have indeed given us much to project with. Even without new advances in AI, present ML techniques have started giving reliable insights given enough data. The scale of data on the cloud has increased massively- thanks to the cost savings the cloud offers. Gargantuan datacentre with embedded tools for AI, data mining and easy deploy for every new patch are already a reality. There appears no reason for this to slow down in near future.

If the AI God , let’s say, is “switched ON” one day, it can be fed enough data to grow in size in relatively short amount of time. In December 2018, Deepmind announced that its AlphaZero was capable of learning every game by itself and perform better than all human players in each. As an example, AlphaZero defeated Stockfish, the smartest Chess playing engine (which itself is smarter than any human player) by learning the game for just 4 hours. This feat is especially remarkable because the machine achieved it without feeding in any initial opening moves or configurations into the algorithm. Most chess engines rely on humans to hardcode the opening books and configurations since the initial possibilities are too diverse . To put it into perspective- AlphaZero would have been able to defeat Kasprov very easily in just 4 hours after it knew what “Chess” and its rules were. That’s 35 years of a great human mind being outsmarted by 4 hours of a computer mind. Now imagine all games and puzzles that you ever played and realize that AlphaZero can do it better than anyone in the world in a matter of minutes or hours at max.

The other school on AI and its future refuses to accept that consciousness can be derived merely from information processing. Roger Penrose argued long back in his book Emperor’s New Mind about the quantum nature of the brain and how consciousness seems to arise out of quantum interactions of the nervous system. While work on Quantum Computers is showing progress and IBM is geared towards making quantum computer susable in some years, most of what is known about AI doesn’t directly transfer to quantum realm. Present day AI is focused on information processing to generate intelligence. If the nature of consciousness, however, turns out to be biological and not logical, it may turn out that we have been beating around the wrong bush. Chirstof Koch known for his work on neurology believes that consciousness, like mass, is fundamental to the universe. He argues that simulating consciousness on silicon and passing the turing test is much different than experiencing it first hand[]. If consciousness does turn out to be biological or chemical, it could be a matter of relief to humans. The knowledge that the smartest being on earth is not sentient still doesn’t dethrone humans from exalted position of “experiencing” things.

While the debate on consciousness remains unresolved- as do many questions in AI- it seems clear that the path already taken will lead to major changes in the world in next 100 years. The smart phone today is getting us ready for that future- both by fetching behavioural data from us and sending back easy result. Technology is already granting us great dividends to the point that 50 years ago, present levels of prosperity, luxury and interconnectedness would have been unimaginable. The promises of AI are too great and its worshipers too many. The cult behaviour of the primal human brain is easily seen amongst all of us. If an AI God/Master is offered to such society, it may well be that blind following will grow around it. There is high chance then that there may be no going back once we do invent a super-intelligent machine there won’t be any going back- not just because of dependency but also the primal need for subjugation of the human mind.

Ray Kurzweil had a positive outlook towards AI in his book where he imagined a universe where intelligence radiates out of the solar-system and saturates all of matter. While that makes us appear to be the pioneer species in a desolate universe, it is also the best case scenario. Anyone who has dealt with uncertainty understands how lucrative and fatal best-case scenarios are. The questions that AI raises about ethics need to be answered by society through effortful and scientific grit. The people must study and understand the principles that govern present day technology and the trade-off between future promises and risks.