Last Updated on November 10, 2020 by Sean B
Facebook might not be the biggest innovator in the tech world, but they do eventually catch up with what their competitors and sometimes they even come up with refined versions of what others are doing. A classic example of that is introducing the concept of stories in WhatsApp, Messenger, and Instagram. Such is the case with the Facebook M Chatbot. When the chatbot boom started in the early 2010s, Facebook didn’t want to be left behind.
Introducing Facebook M Virtual Assistant
Facebook introduced their very own chatbot they called M. Facebook M was more of a virtual assistant than a chatbot and was announced by Facebook in August 2015. It was finally introduced in April 2017 when it was made available to 10,000 users in the US state of California. This bot could be accessed through Facebook Messenger. The concept was to create a chatbot capable of purchasing things on the user’s behalf, arranging gift deliveries, reserving restaurant spots, and arranging travel. M was taken offline by Facebook on January 19, 2018.
M worked by scanning the users’ chat for keywords and then responding with solutions for the situations. For example, if a user reminded their friend that they owe them money, M would suggest the user to use Facebook Pay to transfer the money. It also suggested users do things according to the prevailing time and any upcoming events. It would suggest users send flowers to someone on Valentine’s Day or to buy gifts for family on Christmas.
The Chabot’s Human Supervised Artificial Intelligence
M was not totally autonomous. It tried to tackle the situation the users threw at it, but when matters went beyond the capabilities of the chatbot, a human intervened to straighten things up. It didn’t only make things go right but also helped M grow its abilities and become better by the day.
In this model, human supervision is provided to the AI-powered chatbot up to a specific extent, and after that, the program becomes powerful enough to tackle situations and grow its intelligence and ability on its own.
User Response for M
Facebook designed and deployed M with quite high hopes and wanted it to be a top player in the Chatbot industry. The response of the users was not a very welcoming one. People preferred to use their phones default personal assistants like Siri, Google Assistant, or Bixby instead of engaging M to help them at the time of need.
However, when Facebook finally took M down in January 2018, they described it as an experiment and said that they had only deployed it as a beta version. In any case, the core purpose of M was to increase the economic activity happening on Facebook and ultimately making money for the company, and that did not happen. That caused the people at Facebook to take it down, and according to them, they used their experience from M to create other AI powered products for many of their platforms.
Another reason for the discontinuation of M was that it was an AI product that worked with human aid. The manpower required to keep M backed up could not be justified keeping in view the business it brought to Facebook. People only used it to place calls or remind them to do something. The business part of the bot was never used by the customers, as Facebook expected, and it proved to be the final nail in M’s Coffin.
It was not only Facebook; the users also had a lot of high hopes and expectations for the bot. Once M was capable of pulling off a task, the users tended to throw a harder task at it the next time. And this resulted in people asking M to perform things practically impossible for the AI and machine learning of that time.
What did Facebook gain from M?
Facebook did invest a lot of time and energy into Facebook M and that was not just wasted with the discontinuation of M. During the time M was being used by people, it collected a lot of data from them. The data was later utilized by Facebook to improve their AI research. Facebook put this data to improve the natural language processing techniques they had.
In one of the statements about M, Facebook said that they got an idea of what people expected and needed from a personal assistant. They further said that they would be using the data they got from Facebook M for further research in AI and that they were satisfied with the way M and M suggestions worked.
Facebook and Chatbots
M was Facebook’s chatbot that got discontinued, but that doesn’t mean Messenger does not have chatbots. As of now, there are above 300,000 chatbots on Facebook Messenger, and the number is growing. These include humble chatbots ranging from the ones that tell people the operating hours of a business to the ones capable of closing sales deals on themselves and providing necessary customer support when possible. Facebook Messenger chatbots are considered useful by a lot of companies. As, they use them to reduce the number of customer service and marketing agents they have to hire to get the job done, which the bots can do with a little starting price and not recurring expenses.
In the early to mid-2010s, there was a race for personal assistants, chatbots, and every big player from the tech world wanted to have their own version of a chatbot or personal assistant. Facebook created the M chatbot in that era. It was supposed to give people suggestions according to the circumstances and to tasks that other chatbots of the time were not able to do. However, the whole thing, later described by Facebook as an “experiment”, did not work out as good as Facebook wanted to be, and it was not up to the expectations of the users either. Facebook shut it down on January 19, 2018, but according to them, the data and experience they earned from this “experiment” was used to develop an in-depth understanding of AI and on improving further AI projects of the company.