So by now, most people will have heard enough about chatbots to understand what they are. Recently, since Facebook released their Messenger Platform and interest in focusing on chatbot development, there has been a surge of interest in the world of automated bots. With the help of Slack and Microsoft Bot Framework among others, the technology has really started to take off. What an amazing time to be part of such a new technology! Well actually, it is not a new technology, it has been around for many, many years, and most of the new generation bot developers know nothing about its origins.
A bit of history…
Chatbot development actually goes back to the 1960’s with ELIZA (1966 to be exact), or if you want to go further back to Alan Turing’s conception of the Turing Test, the 1950’s, and I am sure there are other examples in between those years, there are definitely many after such as ALICE and Mitsuku (Winner of multiple best AI awards in the Loebener prize) among many others. There were many different programming languages that have been used to program chatbots over the years: AIML, Chatscript, RiveScript, Java, VB, C#, Python, the list goes on, not to mention machine learning knowledge/natural linguistics interpretation skills etc. I myself have been involved in chatbot and AI development for a number of years now well before the current chatbot craze, I have created my own implementation of the AIML language and interpreter — TAIML which used to power my early days assistants, and have been involved in AI medical applications, personal and business assistants, security systems and many other types of projects that are built around AI and machine learning, including an AI E-Commerce system which I have recently teamed up with the API.AI team on. In the new chatbot era it is no longer a requirement to be a developer that has any knowledge in the above mentioned languages or skills, or for that matter, any requirement to be a programmer. This issue has some good points that I support as a developer, and some points that I dislike.
The good and the bad…
So let’s get the good points out of the way. As someone that has dedicated many years of his life to helping people learn about modern technologies and trying to bridge the gap with the general public and tech, one of things I like is that with platforms like Chatfuel etc, the general public and people with no experience in development can get a taste, and hopefully an interest in taking it further to understand how the underlying technologies work. The reason why I began TechBubble, and also the reasons why I became teaching assistant for MIT’s online computer science courses, is that I fully support others learning about technology to enhance their capabilities. One of the biggest hurdles I have faced over the years, and still do, is getting businesses to understand the importance of utilizing modern technologies as a whole, whether it being through lack of interest or lack of funds, so I fully support platforms like Chatfuel and think they are a great way for non-developers to begin to get an interest in modern technologies, by experiencing their capabilities first hand with no previous development knowledge.
On the flip side of this positive there are some major negatives. Through no fault of their own, these platforms have helped a darker, more deceitful generation of developers emerge. As most of the platforms will utilize some sort of machine learning for natural language processing, suddenly we have a world full of new AI specialists, that in practice would not know where to begin if given a machine learning problem. As someone that has spent a great deal of time and effort learning about AI development and machine learning over the last few years, this issue is one that upsets me. Again, for people interested in learning what chatbots can do, taking their first steps towards learning how to program their own chatbots, or businesses that want to stay up to date with technologies but cannot afford to hire a specialist developer, these platforms are great, but when you have people that have had a bit of experience with them claiming that they are developing, or have developed the next stage in AI, this is morally wrong and affects the credibility of Artificial Intelligence.
Another problem that I have noticed during the last few months, and this relates to the part of the title: “Why the NEW chatbot era is not as innovative as most like to think”, is the way the technologies are actually being used and what type of applications are emerging. Now before I say this I will stress that I am fully aware that there are still many great developers out there, there is still amazing advancements being made thanks to the new era of chatbot development, and that my views in this article do not reflect on the entire chatbot community. Throughout my life in AI/bot development I have focused on projects that could make some sort of a positive difference to the world around me, as have the people that have mentored me, worked with me or I have looked up to in the industry. What seems to be happening at the moment is that the chatbot industry has become a gimmick, I see less and less examples of innovation and more and more “Selena bot” projects, projects that can have no real positive impact on society and that prey on the naivety of it, rather than utilizing the technologies for enhancing the human capabilities or surroundings. This is bad in many ways, but mainly because it is devaluing the whole industry, what is making it worse is these types of applications are being lapped up and their developers being praised for something that took a few clicks and some hours collecting sentences. I was recently having a chat with a fellow respected developer Andrej Petelin, and he said something that absolutely hit the nail on the head:
“Bot builders seem to be mostly about assembling magical lego blocks, don’t they!”
Wrapping it up…
This article is no doubt going to ruffle feathers, but I believe it contains some home truths that needed to be said. The new generation of chatbots can go one of two ways, it can slide downhill or we can raise the level and quality of applications being developed, and utilize this technology for something that positively benefits our existence and the world around us. Lets not kill the concept before it actually takes off!