THE ALGORITHMIC CITY 2035
Visualise this. A series of mathematical entities are now running your city. What started off as finite equations designed to achieve simple tasks have now complicated themselves and self-organised. They were created to simply perform at higher speeds, but over time they became increasingly self aware. This remarkable evolution was achieved in record time since they were able to iterate and evolve at a speed 4 times faster than the Human brain, holding 10 times the data, as in the case of the K Supercomputer from Fujitsu. The designers who set the algorithms in motion themselves now have a diminishing idea on how the equations accomplish what they do.
We are left holding the pieces of a distorted puzzle many millions of cycles after the algorithms have computed the vast permutations the puzzle demands. As you can imagine, this scenario is troubling in multiple ways.
Interestingly, the grammar of Algorithmic repetitiveness mirrors our own concept of Truth, where the same lie repeated over and over gains credibility over time, and gets ossified as a cultural truism.
Our own building blocks, our DNA are strangely Algorithmic in nature, with multiple redundancies and iterations built into the process of replicating. From processing Facebook news feeds to commodity trading, from selecting what video plays next on your playlist to telling you how much money you have in your bank, from emerging uses in Healthcare and selecting dream teams for sports, Algorithms are now the building blocks of our computational experience.
Algorithms have Character. Algorithms have Personality. Algorithms have Soul.
Recently, an Amsterdam based advertising agency curated a unique project. They used a combination of facial recognition software and 18 months of processing visual data from all of Rembrandt’s 346 known paintings to create a unique database. This data then was fed into a 3D printer, which used 13 layers of paint-based UV ink on a canvas to recreate “The Next Rembrandt”, an algorithmically generated artwork which, with a little help from Microsoft and over 148 million pixels, reproduced a completely new painting replicating Rembrandt’s unique style and brushstrokes.
The Next Rembrandt
Another stunning experiment created by Refik Anadol Studio in collaboration with Google’s Artists and Machine Intelligence program, and installed at the SALT Gatala, Istanbul, Archive Dreaming is a 6 meter wide circular installation that employs machine learning algorithms to search and sort relations among 1,700,000 documents. The visual data is sifted and documented by algorithms, but the most incredible thing happens when the AI “hallucinates� when not in use. In this state of deep learning and incidental computation, alternate histories and connections are created using the existing archive that are almost unimaginable to human minds.
Archive Dreaming, Refik Anadol Studio
What happens when an Algorithm hallucinates? Can Machines dream ? AlphaGo, the Algorithm that recently defeated Lee Segol used deep learning and a dream-like higher order state to create it’s own understanding of the game, perplexing experts. We can only piece the gameplay together in retrospect, and that too is not completely clear.
The unbelievably high speeds of computations mean that this is a world we are only dimly aware of and it is humanly impossible to monitor and regulate algorithmic activity. The only warriors in this world are other Algorithms, introduced as checks and balances to spot irregularity and anomalies. High frequency mutations suddenly result in huge blips on our screens, only detected many cycles after the main event has occurred. We can safely extrapolate a future where these all-pervasive Algorithmic entities have a representation in most of what currently comprises the human experience, with continuously eroding lines between what it means to choose as a human, and to choose with augmented abilities.
The Hardcore Algorithm
What were initially considered personality-less bots with great precision but no values, have now started mirroring society’s biases and prejudices.
What happens when Algorithms have personality? Do Algorithms mirror society? In an increasingly divided world can these mathematical entities insulate themselves from our politics ? Our society is far from perfect, and if our own racial biases are any indication, there's no reason to believe the technology and mathematics we’re creating would follow a different ethical system. The sheer speeds of evolution and adaptation of these Algorithms make any potential blunders very scary.
If this wasn’t disturbing enough, it seems that Algorithms tend to the hardcore, progressively moving on pre-decided paths, constantly profiling their users and refining searches in ever tightening circles. On March 23, 2016 Microsoft released Tay, an AI powered chatter-bot to Twitter. Tay was designed to mimic the behaviour of a 19 year old American girl and learn from interactions with other human users on Twitter. Soon however, the bot began releasing racist and sexually charged messages, mimicking the deliberately offensive behaviour of other users. Within 16 hours of it’s release, Tay had tweeted 96,000 times, spouting venom against a wide variety of users, including weird snippets like “Ricky Gervais learned totalitarianism from Adolf Hitler, the inventor of Atheism”.
Is the tendency to become hardcore part of the “moral code� of code? Algorithms learn from their own experiences to refine their actions, but the increase in pinpoint precision comes packaged with a certain myopia. Human brains are able to create and learn from epiphanies, those little understood yet game-changing revelations that allow for massive neuroplasticity and re-adjustments to our world-views, often for the better. Is that too much to expect from our Algorithmic offspring ?
The Algorithm’s tendency to assimilate and magnify in ever-faster cycles lies at the heart of our computational paradigm. If more algorithms are observing more human behaviour on more social media channels than ever before, there is a chance that our most deviant behaviour gets noticed and absorbed, even magnified.
Attitudes of cynicism and racial prejudice take lifetimes to develop or purge. It takes a generational cleanse to overcome colonialism and dysfunctional politics. In the Algorithmic world, manifestation is instant.
The ever faster whirlpools of computation allow deviant behaviour to manifest and take root, sometimes at the cost of millions of dollars, as in the unexplained flash crashes of stock markets, or in the venom spouting AI and Bot experiments. Humans have large reserves of latent cruelty, especially when under the umbrella of authority, where unspeakable acts of cruelty have been committed by otherwise very normal people. We are a complex species, and Algorithms show us what end is in store.
Where do we go from here?
In our emerging world, Algorithmic reliance seems implicit, and its speeds are only increasing. The manner in which we manage the risks of engagement will be critical. If algorithms mirror the biases of societies, it becomes imperative that the groups that conceive of and write this code are open, inclusive, widely representative and diverse, allowing multiple ethical and societal filters to work their way into our mathematics. Eloquent code must become Inclusive code.
Algorithms are mostly unintelligible to a large part of the population, but that paradigm must give way to more openness and closer scrutiny. As the world around us rapidly adapts to an Algorithmic backbone, more of us will need to actively engage in Data Science and Mapping, forming inclusive ethics committees and open learning systems with global jurisdiction. From Governmental to Organisational levels, algorithmic equations that enter our future workplaces must be subject to a high degree of monitoring and scrutiny. It will take the best minds of the future to draft those policies.
At a purely individual scale, policies around Digital privacy, Digital Veganism, and the Right to Disconnect must find voice in the larger world of Algorithmic convenience. Educational programs and playful experiences around this new Mathematics is the need of the hour, in order to demystify it and make it accessible to all. India is uniquely suited to nurture this vision, with a young workforce, digital ready infrastructure and a history of competence with servicing global Infotech needs. If we emerge as leaders in a new way of looking at opaque black-box Maths through our rich and varied lens of inclusivity and celebration of diversities, there's no limits to how meaningful we can make this new Future.
We should be proud Centaurs, boldly augmented humans with Algorithmic resource enhancements, blessed by superior decision-making abilities to chart a bold new world through the jungles of Math that surround us.