By | Abhijit Bhaduri |Keynote speaker, Author and Columnist
She heads HR for one of the largest banks and now has a difficult mandate to execute. She has had to shut down a branch in a small city because that branch is no longer profitable. The reason is automation. Many of the jobs in the bank that were once being done by humans have been handed over to robots or “bots” as they are called.
“Is Artificial Intelligence a job killer”, I ask her.
She protests. “The bots have been programmed in-house in our innovation lab. The innovation lab has added jobs.” She tells me that we cannot run away from the future. The future of the bank is bots doing all the work and being supervised by a single person.
You can bank on AI
AI taketh with one hand and giveth with another albeit with a caveat. The jobs that are being automated often lie at the lower end of the skill pyramid. Those that need basic education and cognitive sophistication.
The jobs that tech is now creating are plenty. But they all need higher education and in depth expertise. As drones become the default delivery model, there will be a need for people to design drones, manage fleets of these, ensure accuracy, keep them light and retrievable, prevent them from hacked. All these jobs will need people who have a certain education, skill set etc that an average person may not possess.
People need banking but they do not need banks. This is not a new phenomenon. ATM machines replaced humans in some jobs but the efficiencies enabled job creation in other parts of the bank. ATMs are more efficient and perform more transactions. So banks need fewer tellers in the branch and lowers operations costs. This in turn enables banks to reach a bigger market and add more branches but with fewer tellers per branch.
AI is everywhere
90% of healthcare data is images generated from X-Rays, CAT Scans and MRIs. IBM’s Watson is now being trained to read images. In April IBM announced a partnership with Apple and the insurance giant Japan Post to try and improve the lives of thousands of senior citizens. AI will do it better than any human. That is because the machine can scan through the vast knowledge base of diagnostic tests of millions of people, choose from almost 800 forms of possible treatments of cancer and sift through millions of test results of drugs and their side effects. The machine does all this in a fraction of a second.
Imagine the difference it could make to your doctor to have access to this level of diagnostic support. That is the same as a researcher being able to do a Google search for any information any time.
Deterministic vs Probabilistic
The world of work that is deterministic depends on clear cut rules. Most algorithms are based on these clear rules. If you encounter situation “A” then do “B”. The “If-Then” component of most jobs will get automated. To stay competitive organizations will adopt technology. Bots will do these jobs better, faster and be much cheaper than humans.
Then there are jobs that are probabilistic. Deterministic jobs have clear rules which when followed leads to the one single destination. The scientific calculator is a good example of machines supporting humans in doing deterministic work.
Artificial Intelligence can supplement human judgement and make us more effective. Imagine being able to match people to jobs that maximize their potential. Or to help HR organize the most effective teams or create new organizational structure that maximizes employee engagement.
Supplementing human talent: brawn, brain and heart
The International Association of Athletics Federations (IAAF) records tell us that the men’s pole-vault world record set in 2014 at 6.16m is 53% higher than the 1912 record. It was in the ’60s when wooden or steel poles were replaced by more flexible carbon-fibre ones. When humans work with machines they can augment their capabilities.
We don’t feel threatened when a crane does the heavy lifting. We don’t cry foul when the search engine spits out the answer to our search queries. We love it when the turn by turn instructions allow us to drive to unknown destinations with confidence. When we tag our friends’ photos on Facebook and Google, they use AI to recognize the face of a person accurately every time by scanning millions of photos. The police are using the same technology to identify criminals. But a despot could use this technology to suppress dissenters. The problem is not with technology. It is with the human who is using the tech.
Talking to machines is slowly becoming common. Apple’s Siri and Google Now are good examples of how AI is being brought in to every day usage. Google has acquired Deep Mind, the company that defeated the world’s best Go player. Deep Mind has built the capability to generate speech which mimics any human voice and which sounds more natural than the best existing Text-to-Speech systems, reducing the gap with human performance by over 50%.
AI, machine learning, and image recognition are being used by advertising agencies to hyper-target consumers by learning their interests and preferences. Facebook uses AI to narrow target segments for their ads. Since all the ads are around your interests, there is a high probability of you getting tempted enough to buy.
In Georgia Tech when a professor added a bot to become his Teaching Assistant, he was able to answer each student individually because most students usually asked the same questions every year as the course progressed. The best part is that the students did not even realize that their questions were not being answered by a human.
The same technology that makes life simpler also puts someone out of a job. But we all like the freedom to withdraw cash from an ATM any time at our convenience. We won’t turn back.
When the AI based virtual assistant in our smart phone helps us choose a restaurant or send a text message we enjoy the moment. We don’t want to turn the clock back to a time when we did not have AI based systems recommending to us what we never knew we wanted to buy. The machine is watching us and learning each move we make. Instead of augmenting brawn, machines are now augmenting our cognitive abilities. Understanding emotions of self and others will be the next frontier. When people learn to work with machines the possibilities are endless.