Rise of the Machines: Deep Learning, Machine Learning, AI, and Big Data

Interested in Becoming a Member?

An SIM Membership like no other, provides you with an abundance of tools, resources and opportunities to help you achieve your professional and personal success at every step of the way! Be part of our learning community of more than 34,000 corporate and individual members.

For more information about membership, please click here »

Member Login

If u are a subscriber, please use ur subscriber login.
If you are a SIM Member, please use your SIM Membership login.

Forgot your password?

Member Login

Forgot your password?
login  Cancel

Sign Up

If you wish to sign up for a SIM Membership, please click here


If you wish to subscribe to Today's Manager, please click here

If you wish to subscribe to Singapore Management Review, please click here

Website maintenance notice: Website will not be accessible from 27 June (11 pm) to 28 June (9 am) due to scheduled maintenance. We apologise for any inconvenience caused.

Home > Articles > Rise of the Machines: Deep Learning, Machine Learning, AI, and Big Data

 Rise of the Machines: Deep Learning, Machine Learning, AI, and Big Data

Bertrand Leong | Today's Manager
June 1, 2018

Deep learning (DL), machine learning (ML), artificial intelligence (AI), and big data hold the promise of a dazzling future; while for others, they may be a source of anxiety, and even fear.

The “singularity” or “Skynet” is the main antagonist in the Terminator movie franchise. It is a fictional neural net-based conscious group mind and artificial general intelligence system that threatens to take over and eliminate the human race. But as lines between fact and fiction blur, perhaps what we thought impossible is now becoming a reality as tech advances in deep learning (DL), machine learning (ML), artificial intelligence (AI), and big data show eerie similarities. These are terms we often come across today, perhaps without understanding their full importance. For some, they hold the promise of a dazzling future, while for others, they remain a source of anxiety, and even great fear.

Big data refers to the use of analytical methods to extract patterns from data that improves performance and helps in decision making. AI is the construction and operation of artificial agents that can solve problems, build knowledge, act and plan logically, make decisions under uncertainty, learn from past data, and perceive and act upon their environment. ML is software that can learn directly from data and improve their performance using more data. ML is part of AI. DL is one kind of machine learning that used complex, multilayer network structures to learn from data.

Director of ML at SP Jain School of Global Management, Dr Debashis Guha has been working on ML-based demand forecasting, AI-enhanced CRM software, and AI-based high-frequency trading says that the impact on all businesses will be epochal and as revolutionary as the industrial revolution or the invention of electrical power. “They will be absolutely pervasive in the long-run, like electricity,” he says. “Even in the short-run, they will be taken up by the tech sector, in finance, parts of the healthcare sector, and the marketing function in any business.”

AI/data Technology
“ML, and especially DL needs plenty of data to achieve good results and thus big data is necessary for ML. Conversely, big data analytics uses ML tools to carry out analytics, so that these technologies are complementary, and we call them AI/data tech,” says Dr Guha.

AI/data is already used to sell products, diagnose illnesses, detect credit card fraud, decide who gets a loan, recommend books or movies, manage election campaigns, and more. Some uses of DL include image processing that is used for diagnosing illnesses from X-rays and ECG images; ML for recommender systems used by Amazon, Netflix, and many others to suggest new pro-ducts to existing subscribers; and AI robot teachers for elementary schools. AI/data will completely change the organisation of work in offices and factories, digital and telephone communications, transportation systems (autonomous vehicles), banking and finance, management of buildings and home environments, and other aspects of our lives.

Research in Pursuit of Human Enablement and Augmenting Technologies
“Research suggests that AI/data can be directed to develop towards more human enabling and augmenting technologies and that we should work towards this goal,” says Dr Guha. In fact, several necessary directions of research are being pursued to incorporate and pave the way for these technologies:

  • Fundamental scientific research to advance the state of the art,
  • Business research on how to implement these technologies and how to reorganise the workplace, and the administrative, management, and strategic structures of work,
  • Engineering research to design the hardware and software platforms that will carry out state-of-the-art algorithms and methods, and
  • Philosophical, ethical, legal, and sociological research on how AI design, legal frameworks, and social structures need to change as a result of these technologies.

To Fear or Embrace?
Strong productivity and income growth and business disruptions in all sectors are possible implications which will arise with these technologies. There will be a loss of old jobs, the creation of new ones, a decline of privacy, a rise in surveillance, and large transformations of governance structures in both the public and private sectors.






Dr Debashis Guha says businesses will
need to evaluate whether their business models

fit into an AI/data world.(GUHA D)

These do not come without their fair share of doubts, fears, scepticism, and anxieties:

  • Fear of the unfamiliar. “This is quite natural,” says Dr Guha, “and will probably go away as these tools become more familiar with time.”
  • Anxieties of large-scale job losses and widespread unemployment. Large-scale job losses will undoubtedly be the case as AI replaces certain job functions. But these losses would probably be partly or even fully offset by the creation of new jobs.
  • Fears that technology may inadvertently lead to an overreliance on tech and human complacency.
    All new technologies have this effect, and human ability has always adapted to doing different things once technology assumed functions that used to be done manually. One should not think of this as a “dumbing down” of humans’ ability to perform simple tasks, but rather as a switch to doing different things.
  • Fears about the militarisation of AI and the development of a “total surveillance” world that could have aspects of “thought policing”. This is a very legitimate fear and states and societies will have to establish the optimal balance between the legitimate needs of law enforcement and national security on the one hand, and personal liberty on the other.
  • Fear of the “singularity” or “Skynet” reality. “This cannot be ruled out, but it does not appear likely in the near future,” says Dr Guha.

Becoming Future-Ready
“More importantly, we need to understand that the genie is out of the bottle and nothing is going to force it back in, until we decide on our ‘three wishes’ from it, which are going to shape our futures. Society needs a generation, maybe even two, to adjust to such a radical change. Our best efforts to adapt can cut down the time from 50 years to maybe 15–20 before society fully embraces these technologies,” says Dr Guha.

He adds that to enable themselves to be future-ready, individuals must realise that many traditional jobs and functions will be taken over by AI. “They will need to reorient their careers and re-train,” he says. “Businesses will have to figure out how to be AI/data-driven and whether their business model fits into an AI/data world. If not, they will have to change their core business model. Even if their core model is not threatened, they will have to do a lot of work to create a new AI/data-driven workplace.”


Copyright © 2018 Singapore Institute of Management

Article Found In

Today's Manager Issue 2, 2018

View Issue

Browse Articles

By Topic
By Industry
By Geography