siliconindia | | July 20179Such is the overwhelming power of AI, and the limitless possibilities of how it can be used for both good and evil, that OECD's Directorate of Fi-nancial & Enterprise Affairs Compe-tition Committee has come out with a paper which suggests that there could be a need to regulate the unfair use of algorithms for firms to collude to fix prices. We are already seeing signs of it when we search for air tickets or hotel bookings. Once airlines and hoteliers have data about your pos-sible travel dates obtained from your search patterns they can fix the price either way.Before we dive deep into AI let's first clear certain fuzzy thoughts about three terms which we will use in this article ­ Artificial Intelligence (AI), Machine Learning (ML) & Deep Learning (DL); these are not to be used interchangeably as there are in-herent difference between them as is seen in the graphic. As we move into the higher levels of the second wave in Machine Learn-ing, we are encountering manage-ment issues that we never thought before. Not only is AI and ML disrupting the workforce, it is trans-forming the competitive landscape by lowering entry barriers for new and nimble competitors, they are fueled by data analytics a key component of AI & ML.Senior managers across organi-zations agree that changes, such as automation, AI, the rise of non-tra-ditional competitors and advances in manufacturing, were imminent, few had conceived strategies for how their organizations might respond. It is sur-prising because mainstream manage-ment thinking assumes that managers reign supreme in determining the af-fairs of their company. The manager, as homo economicus -- the "rational agent" beloved of economists and fi-nance theorists -- chooses the most preferential strategic option from a range of scenarios based on rational calculation of what will achieve the optimal economic outcome. The business case of Algorithmic Competitive Advantage is compelling as the sheer volume of in-formation and data being churned out is impossible for human managers to process and take decisions. Managers are not leading the disruption (primar-ily from technological innovation), but merely responding to it as best as they can. Herbert Simon, winner of the 1978 Nobel Prize for economics, chal-lenged the notion of homo economicus by recognizing the cognitive limita-tions of managers: we can only partial-ly know our options or their outcomes in any given situation. The complexi-ty, turbulence and uncertainty of the contemporary business environment, fueled in large part by more and better information (and therefore transpar-ency), is defeating the contemporary manager because the 20th century mod-el of how we govern organizations has not changed to reflect the realities of the 21st century. It is time now for the AI-Powered CEO of tomorrow! Deep LearningThe subset of machine learningcomposed of algorithms that permitsoftware to train itself to perform tasks,like speech and image recognition, byexposing multilayered neural networks tovast amounts of data.A subset of AI thatincludes abstrusestatistical techniquesthat enable machinesto improve at taskswith experience. Thecategory includesdeep learningAny technique thatenables computersto mimic humanintelligence, usinglogic, if-then rules,decision trees, andmachine learning(including deeplearning)Machine LearningArtificial IntelligenceAmitabh Ray
< Page 8 | Page 10 >