We read a lot of news about chatbots reshaping entire industry sectors by utilizing artificial intelligence (AI), machine learning and natural language processing. Some chatbots are good in assisting consumers in buying tickets or finding good food nearby. Others can keep a simple conversation alive or replace traditional FAQ pages. What most media pundits miss, however, is that such a capability is nowhere near general AI potential, which casts doubts over the very future of chatbots.
Limitations of goal-oriented and conversational chatbots
For start, chatbots are of two basic types: conversational and goal-oriented. You need a goal-oriented chatbot when the ultimate goal is to complete a certain task, say, compare product prices, select the best fit, and complete a purchase. A conversational chatbot can be useful for entertainment purposes but also to make customer experience more dynamic through replacing static pages with a conversational support service.
Chatbots’ functionality to suggest prices or destinations or good food might appear as AI but, if any, this is actually narrow AI or advanced automation based on experience. Narrow AI is an algorithm designed to perform very specific tasks e.g. compare prices and suggest best price depending on a number of other factors as well. It is more of an automation algorithm rather than AI although capabilities to make actions based on voice commands or natural language input may mislead a good number of users into believing this is general AI.
General AI is what most journalists and bloggers mean when writing about artificial intelligence i.e. a machine capable of self-learning, growing its knowledge, and taking independent decisions. Conversational chatbots have some amazing language processing capabilities but they are still far away from having general AI skills either.
As Chetan Dube, CEO of IPsoft, says in an interview with ZDNet:
“If you ask atomic, simple questions, it works. It gives you that impression because its horizontal sweep expands to that extent; it gives the impression that you're talking to an intelligent agent:
Hey, how is the weather?
Can you book me a flight?
Can I get a hotel reservation?
Can I go Chanterelle, a French restaurant?
What is the score of the preseason game between the Knicks and Nets?
All these work. However, ask questions of your claimed chatbot solution. Ask, can it read? Put it to the test of a ten-year-old. Can it read a standard operating procedure for your company? Can it understand what it has read? Can it solve problems from what it has read? Can it create an empathetic connection to my customer based on what it has read?”
What then you use a chatbot for?
According to a report by Gartner, enterprise chatbots are useful in performing the following tasks within the current IT ecosystem:
- Call center help desk,
- ChatOps approvals,
- Equipment diagnostic inventory management,
- Chatbot scheduling agent.
Well, your business can use chatbots to provide help desk and support services, for managing and approving workflows through events generated by a chatbot, for sending out-of-stock notifications, or to schedule meetings on an AI-enabled corporate calendar. All the media hype surrounding chatbots goes far beyond these relatively limited real-life uses of a chatbot, however.
Don’t get me wrong, having a chat-enabled AI without developing an app is quite advantageous for businesses. It can answer most common user questions and replace 1-800 human operators. Nonetheless, it cannot be a substitute for complex queries and cannot provide assistance in complicated scenarios that occur on a daily basis.
At this stage, I see no fruitful use for chatbots outside applications such as customer support, objects comparisons (including pricing lists), or entertainment. Businesses investing in corporate chatbots would better reconsider their strategy with general AI in mind. Sure, developing a general purpose AI is still a thing of the future; nonetheless, investing heavily in chatbots under the current set of circumstances may prove well wrong in the long term.
What is the alternative? Some experts point out at human mimicking agents as an alternative to current chatbots. This is a promising path to take but you still need more advanced general AI to develop such a conversational chatbot.
(Featured Image Source: Mindbowser)
By Kiril V. Kirilov