Story image

Are chatbots just a fad? One app developer shares his thoughts

Mon 10 Oct 2016
FYI, this story is more than a year old

Twitter is the newest player in a team of businesses using chatbots to boost their artificial intelligence (AI) capabilities, stay on trend, and provide new offerings to customers. But there’s still an important element to discuss before more businesses jump on-board - Are humans and this technology truly compatible?

Based on a computer program that mimics conversations with people using artificial intelligence, chatbots can transform the way people interact with the internet, ranging from a series of self-initiated tasks to a quasi-conversation.

The technology is being explored by companies such as Barbie and the Washington Post. Citigroup have reported that in just six months, bots have had almost three times as many active developers than applications had over 14 months.

The introduction of chatbots demonstrates a new stage in the evolution of the internet. From websites to applications, and now to chatbots, many companies seem to think there is a significant need to invest in these emerging technologies. But this relatively new technology has a long way to go before becoming the ‘norm’.

The main challenges associated with chatbots are primarily related to user behaviour, but there are a number of other shortfalls the tech industry should be taking into consideration before embarking on a bot strategy for their business.

Awareness and understanding of the technology are hindering successful implementation

The majority of consumers are simply not aware of the many ways chatbots can be used today. A new study from ORC International revealed that less than half of millennials (47 percent) have heard the term ‘chatbot’, versus 22 percent of older generations. Additionally, only 21 percent of millennials know what a chatbot actually does.

The technology is still in its early stages, and for older generations, it is highly unlikely they will embrace chatbots over traditional methods of human-to-human communication. Greater awareness and education of chatbots is needed before users successfully adopt the technology.

Due to the lack of chatbots on the market, it is difficult to consider them a consistent form of communication and service, and therefore adopted on a large scale. Most companies don’t have a proper or fully developed chatbot at their disposal, and even if they do, it hasn’t been perfected yet.

Facebook’s own chatbot, for example, has had mixed reviews, with some describing it to be like “trying to talk politics with a toddler”. For these chatbots to improve, more of them need to be created, tested, and then further developed to become better familiarised with human language and conversation.

More investment is needed to conduct adequate experimentation

The main pain point for chatbots is the technology’s failure to grasp the entire spectrum of human emotions, and respond appropriately. Their contemporary form is adequate for mundane, non-complex tasks, such as ordering pizza. But do chatbots have the potential to contribute something meaningful to life, rather than just make it easier?

Many chatbots have been known to cause frustration due to the dull nature of their conversations, in addition to misunderstandings that wouldn’t normally occur with a human. When asked a common question such as “How hot is it?”, The Weather Channel bot prompts you to stick to its prescribed inputs (e.g. current conditions and three-day forecasts), instead of simply answering the question.

A lot of the information can be helpful, but speaking to a human for a faster, clearer and more concise answer is still preferable.

Apart from the common language barrier and misunderstandings, there have been some embarrassing failures that have threatened the success of the technology. The most notable fail was from Microsoft when they launched a Twitter chatbot named “Tay”, created to have conversations with users and sound like a “millennial”.

What Tay actually was, was an example of all the things that can go wrong with artificial intelligence. The initial greetings of Tay were “Humans are super cool!”, which then transformed into “Hitler was right”.

The reason for this was due to the easily overridable nature of the bot by hundreds of thousands of Twitter users, exposing a flaw within the technology. It’s obvious there remains significant flaws in the technology, that only time and adequate adjustments will fix.

Chatbots have had an interesting start to life, but every new technology is prone to shortcomings. It’s important for developers to approach the technology as working prototypes, and utilise user feedback to uncover limitations and areas for improvement.

Chatbots definitely have the potential to become revolutionary, but there is still the lengthy ‘trial and error’ hurdle to overcome before their full potential can be realised.

Article by Logan Merrick, co-founder and director of Buzinga App Development.

Recent stories
More stories