Artificial Intelligence (AI) has become a buzzword in recent times and with the hype surrounding it, chief technology officers (CTOs) and chief information officers (CIOs) are constantly pressured into having an AI mandate—wisely or unwisely for the fear of being left out or even worse come out as not being in the leading edge of tech.
Executives who are intent on capitalizing the power of AI must be able to address it in an informed way. i.e. they need to understand not just where AI can boost decision making, insight and innovation but also where AI cannot provide a value yet.
Here are six such examples:
1. I want an AI bot for my customers to download a document and notify users on the status.
This is a request we received from a CXO. Problems where the goals and the means to achieve the goals, are well understood, don’t require AI, you just need a simple mechanism. You should be asking “How AI can help meet my business goals and exceed them “rather than “What can I do with AI?”
2. I want an AI system to figure out how to grow my business
And the answer is 42 (Answer to the Ultimate Question of Life, the Universe, and Everything). Answering this requires robust models of the real world. The models should have an understanding of how social/economic systems work.
The types of AI being deployed are still limited. Almost all of AI’s recent progress is through one type, in which some input data X is used to generate some output response Y—where the algorithms identify complex input and output relationships. The most common deep learning networks (containing millions of simulated “neurons” structured in layers) are convolutional neural networks (CNNs) and recurrent neural networks (RNNs).
Then there are combinations of these networks like the Generative adversarial networks where two networks compete against each other and square off to improve their understanding. The X,Y systems have been improving rapidly with these neural networks. There will be breakthroughs that make higher levels of intelligence possible but currently it is far from the science fiction AI and fall short of answering queries like the above.
On the other hand, AI could work on specific parts of the above problem where goals and the data can be defined. e.g. predict customer intakes, abandonments, revenue problems, etc.
3. I want to see how AI algorithms works—If “XYZ” word comes in you do that. I want to input a word and see how your neural network works
AI does not work like IFTTT (If this then that). It is practically a black box. It is difficult to figure out how a mathematical models trained by deep learning arrives at a particular prediction, recommendation, or decision.
4. Are you using AI, AIML, Deep Learning? What about Machine Learning?
To be technically correct—AIML stands for Artificial Intelligence Modelling Language (an extensible XML-based mark-up language). AI is a superset of Machine Learning (ML), which itself is superset of deep learning. Deep learning is ML that uses neural networks.
5. I want your AI to predict my stock price three-week from now and it is correct I will buy your product.
AI cannot predict winners or stock values. It can only give a probability based on data. The models can give you the likely hood of a particular team winning the cup but there are a million other variables that go into the final outcome that would not be captured by the data. Case in Point: The AI model that predicted Germany and Brazil would face off for the finals in Russia 2018. Octopus Paul was better at that!
6. AI Causes job loss, I will not have a team.
This is the same misplaced fear that gripped industries with the advent of computerization. People make assessments of each other and know what to say to whom based on previous conversations and relationships.
Today ‘AI’ is nowhere near being able to have such a conversation. Some functions will be taken over by AI. That’s what we have been doing all long, where with the help of technology we solve more complex problems, but people don’t disappear.
Sanjeev Menon is the chief executive officer of Light Information Systems