Artificial Intelligence: AI – What will Make it Boom?

I remember watching the movie A.I. by Steven Spielberg back in early 2000. The emotion of trauma brought out by the little boy ahem, the Android was tear jerking, to say the least. At that time, the discussion of artificial intelligence was not a mainstream topic.

Roll down nearly two decades and you cannot read a tech blog without the mention of AI or ML (Machine Learning). I don’t have the numbers at present but I will not be surprised if these two topics currently constitute over 30% of tech news (published in the last five years) out there.

However, it is not only the news. The startups across the globe are continuously working on AI and ML projects. Many of these startups have interesting approaches to solve various problems. Niki.ai, Active.ai, Ikarus.ai are just some of the names, to begin with. Apart from the startups, the larger organizations such as Google, Amazon, and Microsoft also have their AI and ML projects. This raises the question of what will make these projects successful.

The success of AI and ML projects may not only depend on the innovation and the people around it but also the tools that are available. Many of the projects currently use a rule-based approach to automate workflows and processes.

However, to unleash the true AI one must focus on unsupervised learning. That means the system must be able to receive data continuously and process, learn and take actions based on the learning. Can we do it today with little human intervention? Can we do it today with continuous and real-time data?

This is where the big data technologies offer a relief. While there are plenty of companies that are working with big data, there are not many that can actually claim to handle streaming data in a systematic manner. Of course, there are startups that are working in this space as well. Dataxylo, IQLECT, and Vidooly are just a few names in this space.

While many of them are function/domain specific (for example Vidooly is in video analytics and IQLECT is in real-time data analytics) these may be a good feeder into the AI platforms that others are building.

Just building an AI or ML system that takes actions based on historical data may not be the right approach. You would want AI to keep learning continuously and keep experiencing new information to refine its logics. Real-time data handling and analysis may become the key to building a strong AI system or just another run of the mill, rule-based system.

I was recently speaking to one of my friends in a large MNC tech company (one of the three mentioned above) and he seems to be the opinion that today when we talk about data analytics, what we picture is the data visualization. Big Data has been a household name for the last ten years till today we do data cleanup and visualization in the name of analytics. Artificial Intelligence and Machine Learning must not follow the same path. Using data properly, quickly and precisely is important for AI and ML to grow.