Looking forward to 2019 as exciting as it can get for technology, in particular for the broader umbrella of Artificial Intelligence, with chips getting faster and cheaper than ever, and the remotest places and people getting connected to the internet. While anyone would be hardly oblivious to the term AI unless one is absolutely living in a den, we have been consuming the outcomes of AI through many applications and systems we use on a daily basis and this dependence seems to be only strengthening its roots in the coming year. We have been using AI for a while now from relevant google search listing to relevant shopping items on Amazon, or our favorite automatic playlists on Spotify to interactive toys and games, and our favorite assistants, Siri, Alexa, Cortana and Google Assistant, all of which seemed to still exist inside the blurred virtual boundary inside our devices. But Internet of things, Smart Cities, 5G, Quantum Computing, Big Data, and associated technologies have revamped the landscape into a cohesive web of hardware and software, creating a medium ripest for the power of AI to get magnified by multiple folds, the precursors of which are already noticeable with companies using AI from customer to industrial applications. Undercover Recruiter estimates that AI will take over 16% of hiring decisions in the coming decade, Searching for a job online has gone a radical change with AI at backend from having to fill up lengthy forms to relevant customized location-based listings being offered by job portals, From automating record keeping in Healthcare to finding unprecedented connections between pointers in records to getting diagnostics done and recognizing diseases, health care is looking forward to a fundamental disruption which blurs the boundaries between virtual and real. Financial System is another domain where AI has been used to decide creditworthiness of a potential applicant and has since found its way into portfolio management, risk assessment, trading and market analysis and it is seeing no stops in the coming time. Similarly, transportation has been using autopilots for a while now as assisted driving in high-end cars, trains, planes, which is only going to be taken further for granted as the self-driving car technology is being established with the harshest reliability tests. These are real-world examples of having AI at the back of systems which have life and death consequences for millions of people every day. And with advancements in Natural Language Processing, Speech Recognition, Text Recognition, Image Recognition and other core aspects of AI, Augmented Agents, Artificial Customer Care Representatives, Cobots working side by side in our factories, AI Surgeons and Diagnostic Agents are going to be a natural way of interaction with the new world powered by AI.
Ringing bells for the governments and cooperations equally to hasten their foothold on the advancement and perks AI has to offer, while corporations are shaping themselves around AI like Google, Open AI, and mushroomed startups in Silicon Valley, Boston, Montreal and Europe alike, Governments and Public Institutions are speeding up their funding initiatives and branching out to have oversight and monopoly over what AI has to offer, a few of them being MIT recently funding a 1 billion dollar initiative for AI, Dubai government appointing a minister for AI, Saudi government granting citizenship to Sofia (the humanoid robot), Stanford providing the highest remuneration to its AI head, stand as just a few examples reflecting the seriousness with which every small or large scale organization, government and non profits are rounding up their resources towards AI initiatives.
While the seriousness and curiosity behind these initiatives is not just giving into the hype, given we have seen what Deepmind’s Alpha Go, Alpha Go Zero and Open AI Five has been able to pull off, until recently which seemed farfetched and unrealistic to expect from any algorithm, but AI has been able to demonstrate the craft so smoothly leaving even experts to awe. As some more consumer-centric projects, like Self Driving Cars(Tesla Autopilot), Robotic Retail Chains(Amazon Robot Super Market), Robotic Cleaners(iRobot), Bot Assistants(Google Assistant, Alexa), Automated Tax Filling(Turbo Tax) and numerous others have already started to disrupt industries from law firms to agriculture, they only seem to affirm to us what AI is capable of and help us reimagine what future will look like.
While this definitely looks anything far from exciting to imagine AI helping us to cure more diseases, bringing solutions to the water and food crisis, helping us with poverty, helping us tackle sex trafficking, and talking to our kids and teaching them amazing things in a way that they can understand, but for now the reality looks a little gloomy. What is very clear from examples like, AI systems from Microsoft, IBM, and Google failing to recognize African and Afro American people with the same level of accuracy as other communities and in some cases categorizing them as “gorillas”, the search results from google displaying results of “Donald Trump” when searching the word “stupid”, a company “Cambridge Analytica” able to manipulate voters in favor of a certain party, Google Allo suggesting ‘gun’ emoji for a man with Turban, and autonomous vehicles moving pedestrians, is that like any technology AI is as good as the people who use it and expecting a perfection from something created by humans without reflecting the bias that exists between us is like living in a fools paradise. The data that is used to train these AI systems are real-world information that is filled with inconsistencies, biases, individual decisions, and human weaknesses which AI learns as well as its ability to learn the requisite characteristics.
A number of companies, non-profits and public organizations are waking up to the criticality bias in AI with IBM recently launching the initiative to tackle it and more and more universities and influential people and platforms speaking about it and starting awareness and initiatives to help make AI more consistent and fair. But the bias is not the only problem. Given the scale of access, all these AI systems have over the connected world we are living in and the power it gives these algorithms, the scale of loss is very huge and irreversible. The scale of which even Billionaire Entrepreneur Elon Musk spoke of as a technology which scares him the most.
But with all the power AI has to destroy we as humans still focus on and hope for the positivity and relief it can bring to our world. We are building AI to understand us, to be like us, to talk like us, to look like us. We are building it into something that will work for us and work with us to be a friend rather than enslave us. While we are building interactive toys to play with our children, AI assistants to talk to our old, Intelligent cars to help a driver who is drunk, exoskeletons for people who are differently abled, bionic devices for people who lost organs. At the same time developing algorithms, that can understand emotions, understand hope, increase our faith in this life, and be the humans we seek ourselves to become.
With all the possibilities AI is throwing our way to mitigate our problems and crisis, build solutions for problems which existed all along, and create comforts which never existed, everything AI should do but take what makes us human. Empathy, Kindness, Hope, Faith, Relationships and inherent humaneness which bonds us together and makes us feel alive and gives us a purpose to life, should be the last thing we should allow anything to take from us. Rather with lesser problems, we should become more caring and more respecting of others and with more machines become more cohesive and friendlier with people, but connectedness has taught us otherwise. With more and more people responding positively to the question if they think they are lonely even with so much increase in connectedness.
While working on some kaggle challenge in MIT Stata Center, last year, sitting there for hours without talking to the person besides me working on some other challenge, I found it strange, that we were there for so long without even speaking a word. I told him, “while we are building machines which are more like humans, we are turning into people which are more like machines”.
Hopefully, we will understand it goes through the irreversible scale…