If you read through the human history, especially the last 2,000 years, you will notice that famines, diseases and wars were the three most formidable challenges that we had to face (or as I call it — the trifecta of human sufferings). Our response to the trifecta was mostly to seek safety and pray for our survival. History also tells us that when faced with external threats and no possible solutions in sight, we turned inwards and sought solace in numbers to overcome some serious challenges to our species’ survival. Gradually, these cycles of sufferings led to emergence and prevalence of the unifying behavior of collaboration amongst humans. Tremendous efforts and resources, collaborated at extraordinary scale, were thrown at solving the trifecta of problems with some amount of success.
Then about 600 years ago, we heralded the age of renaissance and humans’ collective attention turned towards observing the natural laws and developing “the scientific temper”. Over the next few centuries, we went through three phases of industrial revolutions before our civilization could say ‘cheese’. This science-fueled progress ensured that we solved the trifecta of human sufferings to a large extent. We have now started to look beyond and further than the trifecta and the ongoing fourth phase of industrial revolution is helping us in this exploration. The fourth industrial revolution is arguably more exciting and impactful because of tremendous technological tools at our disposal and the resultant windfall gains for us. Artificial Intelligence (AI) is one such technology powering the fourth industrial revolution which has raised more hopes (and few eyebrows) than others. The story of AI, so far, has been full of ups and downs; filled with moments of celebrations and heartbreaks. It’s possibly had a more serpentine plot than Games of Thrones but it kept on advancing and along the way crossed some important milestones before it arrived on the stage. But AI is here now and it’s here to stay. So let’s understand where we are in the AI journey, where we are headed and what possible choices we may have to make as a civilization in the not so distant future.
Where we are?
AI has been lurking in academic circles and researchers’ labs for a while now but any real progress was limited by the required computing power. Things changed over the decades with the advent of powerful chips and remote computing infrastructure. The watershed event in the growth of AI occurred in March 2016 when DeepMind’s AlphaGo won against the reigning world champion Lee Sedol in the ancient Chinese board game called Go. This was a significant development in the evolution of AI as it paved the way for advanced self-learning systems, which can adapt themselves as they encounter uncertain conditions and unforeseen scenarios, much like human intelligence.
AI as inductive interpolators
In the last four years, we have seen AI taking several strides on its way to become the most significant technology of our time with the potential to hugely disrupt the future. We have seen the emergence of deep learning algorithms like AlphaGo whose underlying technology is a form of general purpose AI that utilizes deep neural networks (similar to neural structure in human brain). Deep learning models enable multiple iterations of reinforcement learning that improves its performance with each iteration. Another case in point is the recent breakthrough of OpenAI’s GPT-3 model. If you have been following the news around GPT-3, you would have heard of multiple applications that researchers and enthusiasts are developing. GPT-3 has probably the most advanced linguistic capabilities amongst its peers and we are still discovering its full range of skills.
Human — Tech Augmentation
The other major area of interest for AI community is to look for ways to enhance human capabilities using AI and find the answers to some of vexing issues around the human longevity. Companies like DeepMind are doing exemplary work to diagnose some of the common but deadly diseases like cancer. They are pushing the boundaries of AI capabilities in the hope of discovering the cure for diseases like Alzheimer’s which traditionally have had an adverse impact on quality of life especially amongst the senior members of the human population. There are also companies like Neuralink, who apparently have lofty aspirations like augmenting the functions of human brain by implanting tiny electrodes in the brain “that may one day upload and download thoughts.” Neuralink and its peers have the shared objective of leveraging AI to further the capabilities of human and effectively change the definition of what’s been traditionally understood to be the limitations of human body. Then there are others like Boston Dynamics who are leveraging AI technologies more from practical use perspective and trying to build more purposeful (but somewhat creepy looking) future “workdogs” to automate the labour intensive tasks.
Where do we go from here?
From learning to identify “Cats” to beating the best of human players in complex multiplayer games, AI has come a long way. Even though, it has already been interwoven into various consumer interfaces like Siri and Alexa, the path forward for AI is still not very clear. There is however, some consensus amongst AI community that the goal is to develop a general purpose AI (aka artificial general intelligence or AGI) and eventually achieve “”Singularity” (simply put, when AI surpasses human intelligence). While tools like GPT-3 have arguably shown some capabilities of general purpose AI, we are still a long way off from reaching the goal of Singularity. AI in its current form and shape reminds me of that episode of Mad Men where they installed the first computer in the office that filled the whole conference room and people wondered at the powers of ‘the magnificent beast’. I believe, AI is at a similar stage and the possibilities going forward are endless. Essentially, based on how are you feeling on a particular day; you can pick and choose the narratives of the AI’s future. But, if you were to simplify the narratives then you can look at it broadly from the prisms of two lines of thoughts that currently exist.
One group is led by academics like Shoshana Zuboff who are skeptical that AI can make any real or meaningful contributions to growth of human civilization and forewarn us that if we are to continue on the current path, we’ll end up with some kind of AI-sponsored doomsday. According to Shoshana, we are currently going through the Age of Surveillance Capitalism where “it is no longer enough to automate information flows about us; the goal now is to automate us”. Skeptics of AI feel that algorithms controlling humans will not be using any kind of brute force strategy but will possibly work by targeting our sensations, emotions and thoughts. Our minds will be so overwhelmed that it’ll be difficult to distinguish between the original and fake information (possibly much harder than telling apart Justin Bieber and Kate Mckinnon’s impression of him on SNL).
The other group belongs to people like Elon Musk and Sam Altman who co-founded OpenAI to advocate for the responsible development and use of AI. These proponents support pursuing the path of developing friendly AI in a way that “benefits all of humanity”. They are already pushing the envelope by developing tools like GPT-3.
Apart from these stark choices as to how to proceed on AI journey, there are inherent challenges that may hinder the growth of AI and need to be solved. The remarkable progress in AI, so far, has largely been due to increasingly powerful computing made possible by the tremendous improvements in processing capacity of chips following Moore’s law. But, there are indications that we may be testing the limits of the Moore’s law in the near future. One possible solution of overcoming the current limitations of computing power is by improving the chip design through technologies like EUV lithography and 3D architecture. The other ray of hope is the quantum computing technology that has promised to take away most of the challenges posed to AI researchers by exponentially increasing computing power currently available to handle large data sets and enabling the development of better AI models and accurate algorithms.
What choices we’ll have to make?
“Sapiens don’t behave according to a cold mathematical logic, but rather according to a warm social logic. We are ruled by emotions.” — Yuval Noah Harari
There is a raging debate about ethical aspects of AI and it’s potential to disrupt the future in ways that we may not be able to control. There are skeptics who raise the specter of AI warlords reigning over the human civilization. They find credence for their predictions of an impending AI doom in incidents like the one involving Facebook’s AI tool and more recently few misadventures with GPT-3 model. This is where, I believe, we will have to make the foremost immediate choice. Will we become fearful of malicious actors (human/sentient) taking over this powerful tool and resist the oncoming change? Or will we collectively evaluate and decide that AI is good for us and march onwards? Frankly, the arguments for ethical AI find resonance with the debate in 1950s and ’60s around the pros and cons of nuclear power plants; only the stakes with the debate on AI are much higher. There has been initiatives to counter this narrative of doom and the good news is that there are indications that we will overcome the fears around AI sooner rather than later as it finds more everyday applications beyond powering Alexa and Siri.
There are also apprehensions among many that the AI has the potential to further aggravate the inequities that already exist in the society. One of the reasons for such apprehensions is that despite the astronomical growth in AI, it’s access still has been confined to limited groups of people and businesses. This is largely because of constraints of infrastructure and capital investments required for developing AI models and associated applications. Most people, outside of the limited groups, still look at AI as a black box (in some ways it is) and have a hard time accessing and trusting it much like the automobiles in early 20th century which evoked fear and awe in the masses. These apprehensions will possibly wither away as the accessibility and adoption of AI increases over time.
I believe, the outcome of the debate on AI will be such that we’ll find the ideal path forward that ensures the co-existence of technology and human. There is, however, a possibility that if we are not sagacious, we may find ourselves in an existential conflict with AI. And, while there may be uncertainty about the choices that we need to make but irrespective of the path that we collectively decide to take, it’s certain that AI journey is going to be enthralling and full of adventures. So, please tighten your seat belts and join the humanity as it explores the next frontier of it’s future and continues the relentless pursuit to define and redefine what it means being human in the august presence of AI.
