How can I help you today? – Chatbots for Good
In the 2nd quarter of 2015 something significant happened that quietly changed the way we use the internet: Messaging Apps surpassed Social Platforms in terms of usage hours.
Users around the world are no longer using messaging apps only to chat with friends and family, but also using them to connect with brands, browse merchandise and watch content.
In the last 10 years, our engagement with technology as a means to connect socially with others has increased significantly. Smartphone capabilities have greatly improved in this time, and messaging has become prevalent because it works seamlessly from a mobile platform. SMS, Messaging and Chat apps have come to the fore and these are now Millenials’ preferred method of communication over emails. This could be because there is more versatility with chat – you can send video, live locations and other expressive media that is more difficult to send via email.
This shift in communication methods means we find it more natural to talk to software the way we are talking to our friends and family. Alongside this newfound confidence in tech, the Chatbot has come to the fore. Chatbots (short for chat robot) are virtual assistants who talk to a user through messaging platforms. You may have already had interactions with chatbots in your life. Amazon’s Alexa, Apple’s Siri and Microsoft’s Cortana are all examples of chatbots. You talk to Chatbots like you would talk to a normal person and when programmed with context they can provide relatively intelligent interactions. Chatbots often have a personality which makes a refreshing change and is much more engaging for the user than the more traditional passive menu options on a webpage. They sometimes even have names and voices!
Chatbots are designed to ask and answer questions like a real person would. ‘Machine learning’ is a branch of artificial intelligence based on the idea that systems can learn from data, identify patterns and make decisions with minimal human intervention. A Chatbot – in most instances – comes up with its answers through a bank of predefined scripts and machine learning. If the chatbot reaches a point where it doesn’t know what to do, in some cases it will deflect the conversation, and in others it will pass the user to a real human. In most cases it will also learn from this interaction. Over time the chatbot will gradually gain knowledge in scope and relevance.
In many instances, Chatbots operate using Natural Language Processing (NLP). NLP is a computer’s ability to understand human speech or text. By focussing on understanding how humans communicate with one another, NLP allows a computer to understand and replicate that behaviour.
Chatbots are developed in two distinct types: Retrieval based and Generative. Retrieval based bots work on the principle of directed flows or graphs. The bot is trained to rank the best response set of predefined responses. Retrieval based bots are the most common types of chatbots that you see today, they are often used for customer support. The programmer can decide the tone of the bot, and the user experience, keeping in mind the brand and reputation. One example of a retrieval based chatbot is Eliza. Eliza was arguably the first ever chatbot, created in 1964. Eliza simulated conversation by using pattern matching. This gave users an illusion of “understanding” but the Eliza had no built in framework for contextualising events.
Generative chatbots are not built with predefined responses. Instead, they are trained using a large number of previous conversations and responses (ie data). Generative chatbots also require a very large amount of conversational data to train. If they do not have sufficient “training” the responses might be arbitrary, with bad grammar and may not make a lot of sense. Alice (which stands for Artificial Linguistic Internet Computer Entity), a chatbot that was developed and launched in 1995 is an early example of a Generative Chatbot. You are still able to talk to Alice on the internet today. However, like many contemporary bots, Alice struggles with some questions and returns a mixture of inadvertently postmodern answers and statements.
Fears around Chatbots
At the end of July 2017, the mainstream media reported that two Chatbots being developed by Facebook had appeared to speak in strange shorthand. The media storm that surrounded this case reflected the fear many people have of Artificial Intelligence (AI). The bots were reportedly left unsupervised and developed their own language which mimicked how humans learn how to speak. There is an idea in the mainstream that one day robots could gain equal or superior intelligence to humans.
Contemporary culture reflects this fear of AI in films and television programmes. The film Her (2013) is set in a near future in Los Angeles. The protagonist purchases an operating system with a virtual assistant who has AI human voice. ‘Samantha’ – the virtual assistant in the film is eerily similar to devices like Amazon’s Alexa and Microsoft’s Cortana. The film explores the idea that although technology promises to give us connection and communication, this is just the illusion of meaningful relationships.
Innovation in ChatBots
In 2016 a 19 year old university student called Joshua Browder created a chatbot called DoNotPay which successfully overturned more than 160,000 parking fines. DoNotPay gave free legal aid to users through a simple-to-use chat interface. He has now developed a new chatbot, using Facebook Messenger, that can help refugees fill in an immigration application in the US and Canada, and in the UK it helps them apply for asylum support. The London-born developer worked with lawyers in each country, as well as speaking to asylum seekers whose applications have been successful.
Chatbots for Good
Ulula is an example of a worker engagement technology that leverages chatbots to deploy and analyze at scale automated surveys and two-way feedback, to gain visibility into supply chains to detect risks of human rights abuses. In the words of Ulula CEO Antoine Heuty, “Ulula designs chatbots to engage workers directly and anonymously and get honest feedback about labor conditions to create more responsible and transparent supply chains.”
Ulula’s solution leverages simple communication channels like SMS and IVR (voice calls) in addition to messenger applications and an app in almost any language to ensure an accessible solution. With programs across sectors and regions, Ulula works with diverse collaborators such as the International Peace Information Services (IPIS) to gather difficult to access insights from workers and stakeholders. In 2018, IPIS leveraged the Ulula solution to measure social, environmental, and human rights indicators in Eastern Congo. The analysis underscored the persistence of child labor, forced labor and gender violence on artisanal mine sites and communities.
People all around the world now use messaging apps to communicate with each-other. As a result of this we have become accustomed to talking to people in this way vs other methods such as email and over the phone. Chat allows us to communicate using video, voice and emoji – methods that are more expressive than traditional communication methods. This has made way for the Chatbot in our online lives. Chatbots are often used to identify customer problems before they are passed onto Help desks. More recently they have been used in ‘home hub’ products that you speak to as you would a normal person. The more modern Generative chatbots are seemingly able to ‘learn’ in a way that is eerily human like. This can be used to explain why fears have arisen surrounding Chatbots. The media storm that surrounded a case where two chatbots began to start speaking to each other in code at Facebook HQ, and the film ‘Her’ where the protagonist falls in love with a virtual assistant are both reflections of this. However, recent innovations in Chatbot design have shown that they can be used for good too. Browder’s and Ulula’s chatbots are examples of this and these kind of chatbots can put the power back into the peoples hands and create more responsible and transparent supply chains. This in turn can help to address global issues by providing justice and improve working conditions for people around the world.
By Jess Knights, Creative Marketing Assistant at SustainIt Solutions