Chatbots as a public relations tool: proceed with caution
Facebook launched a chatbot platform for its Messenger application last year. More than 1.3 billion people use Messenger so inevitably the market got hot.
Chatbots are software applications that interact with people in natural language. They’re a new channel for discourse between an organisation and its publics.
100,000 bots have been created on Messenger in the past 12 months. Go and check out the platform and try some of the bots out for yourself.
Bots aren’t new. Internet Relay Chat (IRC), SMS services and automated website chat services have all existed since the early noughties. Automated email has been around for even longer.
What is new is the resource that Facebook is investing in the market. It coincides with the growing sophistication of software or so-called artificial intelligence (AI).
Voice recognition and a new breed of home assistant products is a third driver.
Amazon’s Alexa and Google Home use voice as an interface between human and machine. Both Alexa and Google Home provide native tools to enable developers to build applications for the platforms.
Application of bots in public relations
I’ve identified five applications for chatbots in public relations:
Push out messages in realtime. Used by media organisations to distribute content. Also in crisis communication.
Answers simple requests for transaction information such as environmental data, transport information, or personal records such as a bank balance.
This guides the user through a process using a decision tree. It’s commonplace in learning and development or to automate an FAQ.
Responds to a query by searching a data set for the most likely answer. Google’s search engine is an example.
A learning system that adapts its behaviour based on engagement with a human beings.
You don’t need to be able to code to build a bot although a basic understanding of interfaces, system design, and user interfaces is useful.
There’s a fast growing market of developers and tools.
Staging and testing is a critical part of a new bot deployment. It’s incredibly easy to cause offence.
You’ll have notice how quick LinkedIn and Twitter users are to rage about automated messages sent by some users to new connection.
The opportunity for a conversation to turn sour and the reputational damage of a bot engaging inappropriately in a human conversation could be severe.
Microsoft was forced to pull a two-way AI chatbot from Twitter in March last year after it become a racist, Hitler-loving sex promoter.
The bot was intended to help Microsoft improved its voice recognition software by learning from Twitter conversations. Inevitably Twitter users gamed the bot.
Twitter has a bigger bot problem. The anonymous and open nature of the platform means that automated bots are used to pollute and hijack conversations.
It raises a significant ethical dimension. If a bot creates content that breaches copyright or defames an individual or organisation, who is liable?
Ultimately convincing people to interact with a robot will be a significant behaviour challenge but it’s no more significant than previous shifts in technology.
My view is that we’re at the early stage of the adoption curve of a technology that could prove incredibly exciting for public relations.