It might seem as if words like ‘chatbot’, ‘AI’ and ‘machine learning’ have exploded in news media over the last ten years. That’s because they have and it’s really only in the last five years that the word ‘chatbot’ has become much more widespread and more commonly used by newspapers. Interestingly, it’s only in the last three years that ‘chatbot’ has become a casually inserted term within ordinary articles which are not explicitly about tech or AI: despite the fact that newspapers have been reporting on chatbots since the 1990s.

The word ‘chatbot’ has become normalised in news media from its ‘chatterbot’ origins, and while the frequency of articles which casually make reference to chatbots has increased so have articles which mention chatbots alongside wider conversations about artificial intelligence. Ireland provides a useful case-study of the media and chatbots. The Irish Newspaper Archive is one of the most extensive online archives of Irish newspapers. The archive contains 45 unique newspaper articles that mention the term ‘chatbot’ dating from 1997 to 2019. It seems like a paltry number considering the overwhelming popularity of the topic worldwide, but the data is indicative of an intensifying receptiveness of Irish newspapers to the topic of chatbots. As well as an escalating normalisation of the term in everyday reporting.

An increasing acceptance of chatbot narratives

These 45 articles can be broken down into six categories: where ‘chatbot’ is mentioned in an otherwise non-chatbot related article; articles specifically about chatbots; where ‘chatbot’ is mentioned in an article about AI; articles about chatbots and industry; articles about crime and chatbots; and articles about chatbots in film.

72% of these articles fall into the top three categories of Mentioned (M), Chatbots (C), and Mentioned with AI (M+AI). The majority of all articles occur between 2015 and 2019, a finding which indicates an increased receptiveness of Irish newspapers and news agendas to the topic of chatbots within that time frame. Break down the top three categories from 1997 to 2019, and we can see a sharp increase in usage of the term ‘chatbot’ as well as a greater interest in AI.

In 1997, the Evening Herald published a short article about chatbots in a regular column entitled ‘The Internet’. The article describes chatbots as “virtual personalities” and a handy tool for brushing up on communication skills, and it makes reference to “the chatbot site” as a now broken url. Here, you can practise communication with “multilingual Fred, the overeducated Barry de Facto, and Dawn, whose moods are liable to swing rapidly between extreme shyness and over-the-top enthusiasm”. The column shows its age, as the article above describes the early internet pioneer AOL as a potential sponsor for an upcoming U2 tour.

In ‘Is this the Future of Voice Recognition’ by the Irish Independent in 1998, the narrative is cautiously realistic about the capabilities of chatbots, which the author describes , at the time, as “notoriously bad at pretending to be human”, and when looking to the future they feel that authentic human conversations “are no more realistic now than they were 25 years ago”. While focussing on the convenience aspect, the article describes the key to good “chatterbots”: “to promise help, but not meaningful interaction”.

In the noughties chatbots are given a bit of a hard time from newspaper narratives. They are described as “dumb internet things” with “lofty goals” in 2001, but from 2013 onward we see a shift in opinion towards a more positive, inquisitive narrative. We read about chatbots as tools for law enforcement in luring paedophiles online, the “bright future” of Microsoft’s Cortana, and how chatbots can play a vital role in healthcare. The narratives take a more pessimistic view in 2017, however, as the perceived threat to ethics, humanity, and our workforce plays a more prominent role.

Dystopian narratives VS applications in capitalism

‘We need a Code of Ethics on just what AI is for’ by the Evening Echo explores bias and ethics in AI design, and the need for “a system of morals”. In 2018, we read that the “chatbot revolution is imminent” with “deep engagement” for users (Irish Independent) – a narrative far removed from the call for meaningless interaction two decades previously. Personalities such as Alexa and Tay are scrutinised for their offensive capabilities, and in 2019 a particularly doomsday-ish article by the Belfast Newsletter describes “frightening cases of AI gone wrong” with the threat of autonomous military weapons, and a completely inaccurate analysis of the Facebook chatbot experiment where chatbots learned to ‘secretly communicate with each other’ (they didn’t).

Some of these narratives have fallen prey to technological hysteria and doomsday future telling in one-sided commentary, however the majority of articles published post 2010 assimilate the word ‘chatbot’ into everyday reporting (on movies, motherhood, advice columns, youth culture, and politics), or they make an effort to take a measured approach to its application in society by including critique from experts in the field and university lecturers. Just one article gives an alternate view to the threat of automation, attributing “hysteria over chatbots” to “our natural instinct to credit AI with an awareness that it does not have” (Irish Independent – a year later from its previous doomsday article).

What does this mean for the chatbot industry and bot makers?

Evidence from Irish newspaper narratives around the topic of chatbots suggests that journalists have become more receptive to their applications in society, and more inquisitive about how this technology will develop into the future. Since 2013, news agendas have paid greater attention to the practical applications of chatbots in the world around us and to providing convincing examples of their synergy with industry, commerce, and day-to-day life. The early dismissal of chatbots as “dumb computer things” has been replaced by a willingness to engage on a meaningful level, to explore the positive and negative implications of chatbots, and to provide readers with greater insight as to exactly how they may become users of this technology rather than just observers. For those in the chatbot industry wanting to get the message across, this has to be good news.