The Rise of Machines Over Bulls & Bears

What’s the buzz all about?

Over the years, there has been considerable speculation surrounding the development and use of artificial intelligence (AI). What seemed as a faint possibility has indeed become reality.

Especially in the last couple of years, the use of AI has crept in to many aspects of our lives. At work, home, retail, healthcare, banking, travel and related fields. Stock markets and trading seems to be the next on the list.

AI in financial markets

In the last couple of years, AI has been extensively used in financial markets in the decision making process. In the usual course of business, a stock broker or fund manager keeps a close watch on the different stocks including but not limited to mutual funds, hedge funds, pension funds and the like. Stock brokers / traders may buy and hold stocks for varying period of time. They acquire /dispose off stocks with a view to make profits. Fund managers of mutual funds essentially manage a whole fund of stocks, bonds or other assets that is bought with the investors’ money. They are more like stock pickers who focus on assets to invest in.

Irrespective of whether it is an individual stock broker or a fund manager of a fund house, they indulge in crucial activities of credit evaluation, risk analysis, forecasting, planning and management of stocks/portfolios. This was traditionally done by regular collection, analysis and drawing conclusions on voluminous statistical data.

Human expertise and experience is used in empowering AI systems (software / robots) to make trading decisions. This is done by identifying patterns in data. The more the volume of data, the better prepared the AI systems are to take decisions. As such AI uses computing, language processing and data analytics and this is like fodder for cattle!

How AI is used in financial markets

There are numerous applications of AI across stock markets. A look at some key applications:

  •     Algorithmic trading

    How AI is used

    Popularly known as ‘algos’, they are used in processing large /bulk orders using automated pre-programmed trading instructions after taking into account variables such as time, price and volume.  This was developed to save traders time from sending out routine instructions manually.

    By this, it is attempted to minimise cost and risk / error in executing an order. Useful in high frequency trades (HFTs), algos are widely in banking and Mutual funds.



    Buy 50 shares of a stock when its 50-day moving average goes above the 200-day moving average.

    Sell shares of the stock when its 50-day moving average goes below the 200-day moving average.


    An algorithm program is written based on the above instructions which will automatically monitor stock prices and will buy / sell when conditions are met. The trader / fund manager need to keep a continuous watch on the same.


  •     Voice recognition

    How AI is used

    AI is used to analyse text and human voice. Based on the analysis Algos are developed to trade in stock markets.



    IBM’s Watson is used by Japans financial institutions to analyse texts such a s corporate disclosures and social media.


  •     Data mining

    How AI is used

    AI is used to sift through huge volumes of historical data and identify patterns of behaviour and solve problem through data analysis.


  •     Market surveillance

    How AI is used

    The practices put in place and procedures followed in investigation of illegal trading practices in the securities market. This is done to ensure that markets are in order and traders are assured of credibility and fair dealings.



    Initially used by the London Stock Exchange, the LSE extracted a data set from a market’s order book, normalized the data, stored it in a manner that leveraged the then-emerging big data tools, and began running machine learning type algorithms based on this.


  •     Deep learning

    How AI is used

    A highly specialised technology within AI. Scientists are in the process of creating computer models similar to the structure of the human brain, with the ability to think like human beings. Called artificial neural networks, these systems use a combination of sophisticated algorithms and voluminous data. The data is fed into the neural systems to enable them ‘learn’ to differentiate and distinguish similarities and differences.



    A robot could be used in household chores by being fed with the instructions / opinions of large volume of data fed through AI in order to complete everyday tasks like cleaning, organising, driving a car etc. It could use speech recognition as well. The same could be customised to dealing with financial markets as well.


  •     Machine learning

    How AI is used

    Machine Learning is where computer algorithms are used to autonomously learn from data and information and improve the existing algorithms.



    They can be used to predict stock prices or the impact of a new regulations on the markets etc.

    Insurance firms use it in claims processing.


The above are a few but important applications in the financial markets. Other applications are in the R&D stage. It only seems to be a matter of time before they are out and available.

The story back home

India is not lagging behind.  In Delhi two computer scientists cum entrepreneurs have successfully utilized AI in portfolio management services.  The fund management firm runs on a mathematical model that integrates data analysis and AI to manage their stocks and funds. Hedge funds are increasingly turning to artificial intelligence in order to spot trends to try to make money for their customers.

There are at least 22 Indian AI companies (and counting) that started developing as early as 2005. But the majority of AI companies in India have come up in the last 3 years. The trend is only expected to continue.

Creation Vs the Creator?

It is believed that it took about 30 years to develop and build the first self learning algorithm that analyses and predicts stock markets. Based on machine learning, research is continuously done globally and in India to improvise the use of AI into financial markets. Asset managers have increasingly resorted to using artificial intelligence in managing their investments. Using techniques such as data mining and deep learning, they are trying to analyse tons of data and (including social media news and information) to develop portfolios for investment.

As recent as December, 2017, Bloomberg reported that a hedge fund manager and a computer scientist have found a promising new way to use artificial intelligence to pick stocks over longer periods than the typical machine-driven approaches that are currently being favoured by Wall Street. Bloomberg also reported in August 2016 that a hedge fund robot out smarted its human creator by investing more profitably than him. With efforts on to integrate to bring in the “creative” element into machines, it remains to be seen whether the creation outsmart the creator.


Artificial Intelligence: Laws for the Virtual Intelligence

The concept of artificial intelligence originated way back in the 1950’s. It originated as an idea wherein humans could, at some point, develop machines that could actually think (like human beings).

The father of Artificial Intelligence, John McCarthy, defined it as “The science and engineering of making intelligent machines, especially intelligent computer programs”.

Wikipedia defines Artificial Intelligence (AI) the theory and development of computer systems able to perform tasks normally requiring human intelligence, such as visual perception, speech recognition, decision-making, and translation between languages.

Simply put, Artificial intelligence is a way of making a computer or software or robot think and act like humans do.

Process of AI

Generally, the process of AI involves the study of how the human brain thinks, works and makes decisions. On the basis of such study software systems are developed to “think” similarly.

Development of AI is not just a Science. It is dependent on other disciplines of study as well – computer science, linguistics, mathematics, engineering, psychology and the like. Study of logic, reasoning, decision making, problem solving and judgement form the crux of developing the AI programs.

Areas of AI use

  1. Language understanding, analysis and assimilation
  2. Speech recognition
  3. Robotics
  4. Video gaming / virtual reality
  5. Data analysis / data mining

Pros & Cons of AI

AI governed by Laws?

Advanced nations are yet to catch up with the need to put in place stringent legislations to tackle with issues arising out of ownership, development, use and monitoring of AI.

India is yet to have a comprehensive law governing development and regulation of its artificial intelligence. There are some independent bodies whose aim is to better research and development of AI to its maximum potential. Non-profit Societies such as Artificial intelligence Association of India (AIAI) were incorporated in 2009 with the objective to understand mechanisms underlying human intelligence and integrate it with machines; increase public awareness of AI.  Another organisation – the Artificial intelligence and Robotics – an arm of the DRDO is engaged in developing AI that can be used by strategic government departments and law enforcement agencies. However there is no exclusive legislation in place yet. Development of laws also become imperative in face of intellectual property rights issues with respect to such developed AI machines / robots / software programs. Our existing laws do not charge a machine or robot for criminal charges or IPR infringement. Tapping on the increasing potential of AI, especially after the “Digital India” program, the government announced in the Union budget, 2018 government doubled the fund allocation to the think tank NITI Ayog for development of AI programs. This has clearly been done with the objective of the national defence and security and also facilitate ease of doing business. Human intelligence....to develop artificial intelligence!

In the fortnightly podcast, Mr. A K Mylswamy shares his experience arguing in yet another corporate law case on Black & Green Mobile Pvt vs Magesh Madhavan.
Listen in!



AI Law


Qualifying share

A share of common stockowner by a person in order to qualify as a director of the issuing corporation in a corporation that requires directors to be shareholders.

Ratio decidendi

The ground or reason of decision. The point in case which determines the judgement.

Sale per aversionem

In the civil law, a sale where the goods are taken in bulk,or not by weight or measure, and for a single price, or where a piece of land is sold for gross sum, to be paid for the whole premisses, and not at a fixed price by acre or foot.

A sale per aversion is sale of either distinct or separate immoveable, such as a field enclose or island in a river, or an immoveable property sold by certain bounds or limits.


Limited; abridged; reduced; curtailed, as a fee or estate in fee, to a certain order of succession, or to certain heirs.




Released in 2015, the core of the plot is artificial intelligence.  The police department in Johannesburg use robots as part of its force to control crime. They procure these attack robots from a weapons manufacturing company. A young engineer working for this company designs software to make the robots think and feel like humans. However, his project is turned down by the Company and the engineer uses a damaged robot to experiment his latest software. In an unpleasant turn of events, the engineer is kidnapped by criminals who force him to program the robot – “Chappie” to assist them in committing crimes. As the authorities begin to see Chappie as a threat, they go all out to ensure he is destroyed. Will a thinking robot be the end of mankind? Watch the movie and find out.

Also aired in the year 2015, this television series is a science fiction drama based on the theme of artificial intelligence and robotics. A busy family buys a highly developed robotic servant called a “Synth” to do their household chores and ease their lives…but it transforms their lives forever.


By far the greatest danger of Artificial Intelligence is that people conclude too early that they understand it.

Eliezer Yudkowsky

© Copyright 2018 - A K Mylsamy & Associates LLP


1. Day to day applications

AI can be effectively used in daily needs such as GPS while driving. It can also be found in use of smart phones and their applications. Lady Siri and Cortana are the early birds of AI! Booking travel and stay tickets online, ordering food and groceries and online retailing are all examples using AI.


2. Repetitive / monotonous jobs

Jobs that are routine, monotonous and repetitive in nature can be accomplished by machines with artificial intelligence. There is no room from fatigue and related error. Therefore, the task can be accomplished with greater accuracy and speed. Chances of error is minimised.


3. Minimal (or no) error

Automating routine processes will actually reduce errors rising from various factors such a fatigue which occurs to humans. As a result it increase productivity and takes lesser time.



1. Expensive

the AI software, machines are generally very expensive. The cost of periodic up gradation of software and maintenance of systems / machines cannot be afforded by all.


2. Constant up gradation

Since software technology is no constant and keeps changing periodically, it may be monetarily and practically impossible to keep themselves updated.


3. Restricted scope for improvisation

The human brain’s creative thinking cannot always be replicated or customised by machines or software especially if they are designed to do routine tasks.


4. Unemployment

There is always the threat of large scale unemployment if major work is undertaken by AI machines / robots.