According to Infosys Ltd., Artificial Intelligence is already widely utilized, leading to actual results and helping in the formulation of business strategies in more than seventy-three percent companies around the world. This idea is, however, far from reality according to Boston Consulting Group and MIT Sloan Management Review. According to them, hardly one in twenty companies of USA incorporates AI into its proceedings. Businesses are yet to understand the full potential of AI. IBM Watson Technology from IBM Global services is the only major player in the arena, and its growth is limited by the lack of understanding. Indian IT companies need to develop a realistic understanding of AI and then spread the knowledge among the customers.
The hype and reality of AI
Google’s AlphaGo Zero learned the most difficult board games in three days, and Nvidia’s AI generated photorealistic pictures of good-looking people through pictures of celebrities. News like these creates a lot of hype around AI. The reality is that both Nvidia and Google used generative adversarial networks leading to pitting two AI systems against one another. But before they were pitted, they were coached which meant that both the problems and outcomes were pre-destined. Businesses use too many factors without any pre-defined outcomes which makes these approaches redundant for them.
Hurdles in inventory management and health-care
Inventory management involves maximization of customer services and sales along with minimization of inventory investment. But the rules, in this case too, keep changing and just a change of algorithms do not offer much help. Using AI in assessing health data makes it even more complicated. There are numerous factors like customized treatment and preventive conservation that needs to be carried out in real time.
The limitations to AI
AI uses deep learning techniques which means it sees patterns in the examples that you give it, stores the patterns and tries to find similar patterns in future scenarios. Thus, the more examples you give, the better equipped it gets. But AI does not understand the things that it has analyzed and therefore fails to apply the analysis to other contexts. It has no power to distinguish correlation and causation. The other problem with deep learning is that the patterns it has identified remain unknown which makes the prediction of its responses difficult. Businesses cannot afford such responses and need actual logic behind each decision made.
Ways to a more developed AI
If AI has to function like it is expected to, it needs to include more information sources than the usual computer systems. The problem is that even with all the information, most companies have disorganized data sets. To add to that, there is lack of knowledge about how to put the data sets together. The opportunity for Indian IT lies in the fact that it can facilitate data clear-up, design a new architecture for information, and develop a vision for the implementation of AI. All these are matters of time and experience which means now is the time to begin working on it.