Chatbot Tsunami: The Good & Bad
The widespread use of cell phones and social media has made text-based communication, also known as texting, a mainstream communication method. For example, a Gallup polls hows that texting is the dominant communication method for Americans under 50. Popular text messaging platforms, such as Facebook Messenger and WeChat, all boasted over 1 billion active users monthly.
However, a person can hardly text with multiple parties at the same time, let alone do texting 24/7. That’s where chatbots come to the rescue. A chatbot engages a person in one-on-one text messaging to automate various tasks. For example, Sephora uses a Facebook Messenger chatbot to help users book an in-store makeover, while Kohl’s uses a chatbot to help customers locate a store or find deals.
As a marketing expert pointed out, a chatbot is meant to scale out brand-customer communications and assist customers better. First, a chatbot will never feel tired and can text with hundreds of thousands of users simultaneously to provide them with immediate help. Second, a chatbot will always remain calm and polite to help users even in difficult, emotional situations. Third, a chatbot does not have a personal agenda or ulterior motives and could offer users with objective and fair services.
On the other hand, studies show that most chatbots are just bots with little chatting capabilities. For example, one study shows that 70% of Facebook Messenger chatbots were unable to answer simple questions, and another study shows that it is still quite challenging for chatbots to handle a simple task, like scheduling a meeting. Additionally, these limitations cause “chatbot spam” — using chatbots to simply broadcast one-way messages, similar to email spam, but in the form of text messaging.
Although it may take 10 to 500 years to achieve the ultimate human-machine symbiosis envisioned by Licklider, a practical and urgent question seems:
What can we do today to take full advantage of the good characteristics of chatbots and avoid their bad ones?
My next post, AI Chatbots: Their Good Uses, will attempt to answer the question.
