In trying out the new bots, the launch of which Facebook’s head of messaging David Marcus called “the first day of a new era”, it’s obvious that Facebook and its partners, including news organisations such as CNN and the Wall Street Journal, need to learn a few things, and fast.
User intrigue over what a chatbot can actually do quickly turned into loathing, once they began being spammed through a channel hitherto restricted to personal messages from people they care about.
While users might put up with spam on a social network, messaging systems are more akin to a phone call or SMS, where sender is essentially shouting in the recipients face “talk to me now”. That’s fine if it’s a select friend that you might want to actually talk to, but if it’s a brand you’re very unsure about, it quickly grates.
The trouble begins when you realise you don’t know how do you stop the messages. It’s a bot, so perhaps asking it to unsubscribe in natural language should do it? Not so for the CNN bot at least.
As it turns out a user has to simply send the word “unsubscribe” without any other text to prompt the subscription settings for the CNN bot.
The Wall Street Journal bot was a bit more intelligent. Sending it the word “stop”, as a UK user attempting to unsubscribe from premium-rate SMS messages has learned, spawned the bot’s alert preferences.
Sadly unsubscribing did little to stop the barrage of messages, forcing users, including myself, into actively blocking the WSJ bot to stop the spam.
Even when the bots operate in the way they’re obviously meant to, the interactions have proven hit and miss.
Chatbots are not new. They’ve been around for decades with chat services such as Internet Relay Chat (IRC) having automated services that reply in channel since the mid-90s.
But Facebook’s push for chatbots has opened up new possibilities for brands, publishers and services to reach users who wouldn’t ordinarily install an app. To engage them where they spend their time with their guard down, with friends and family.
Teething problems can be excused, but issues such as these within a space where users are not going to be tolerant of spam and ineptitude could put them off for life.
One day, or even just a few hours, of spamming a user could be enough to earn a complete block – as in the WSJ’s case – and put them on the back foot, skeptical that any other bot will be different.
Chatbots hold potential to be something genuinely useful and interesting, but Facebook and its partners need to sort out their teething problems and quickly, otherwise they could poison the whole chatbot space.
Learning at the expense of users isn’t going to cut it this time, whether you launch a service in “beta” or not.
- Facebook has a new army of chatbots – but what can you do with them?
- Bots are the new apps: Now anyone can build their own version of Microsoft’s racist, sexist chatbot Tay
- What is a chatbot, and should I be using one?
guardian.co.uk © Guardian News and Media Limited 2010