Chatbot-use in governance good, need to have smarter chabots
The government, recognising the potential of chatbots, last week announced that it would be incorporating an AI-based conversational platform similar to Google Assistant or Alexa to provide government services in multiple languages.
The use of chatbots for customer service—first adopted in India by, among a handful of other consumer-facing sectors, the banking and financial services industry—is set to become ubiquitous with the government looking to adopt these in governance delivery. During the start of the pandemic, many states governments launched their chatbots to disseminate information about the virus, and then, the Indian Council of Medical Research partnered IBM to create a chatbot for better management of data services.
While most chatbots have a menu-driven response system, IBM’s Watson, in this case, had also incorporated natural language processing to respond to all kind of questions in different languages. The government, recognising the potential of chatbots, last week announced that it would be incorporating an AI-based conversational platform similar to Google Assistant or Alexa to provide government services in multiple languages. As per news reports, the National e-Governance Division of the ministry of electronics and IT has invited proposals to build the conversational platform for the deployment on the UMANG app. The software will be able to convert speech-input into text and vice versa. Moreover, on Tuesday, while announcing the features of its vaccination registration app Co-Win, the health ministry indicated that the app would also come pre-loaded with an intelligent chatbot to answer user queries.
India’s tryst with chatbots in governance may be new, but they have a proven track record in other countries. In Arkansas, US, the government uses chatbots for more than 900 services. The government bot Gov2Go caters for the needs of most citizens. Kansas has been using Agent Kay, which answers queries on government services and also provides tax-assistance. Agent Kay is also integrated with Facebook to give it wide reach. India, no doubt, needs to take cues from these models and build upon such ideas as large swathes of the country’s population are illiterate and not digitally-savvy. Thus, a voice-activated bot makes sense.
The more urgent need, however, is to make the bots more intelligent. A NITI Aayog discussion paper, National Strategy for Artificial Intelligence, published in June 2018, emphasises the need to incorporate conversationally-smart chatbots. In this regard, the government can learn from the winners of the Turing prize or Loebner prize, who have created chatbots that trick humans into thinking they are conversing with another human. Mitsuku, a Japanese chatbot that won the 2016 Loebner prize, has gotten better over time and can hold a longer conversation. It is also being used to help people with depression. The government chatbots require a similar level of smartness. The government also needs to refrain from burdening citizens with as many chatbots as services and instead integrate more services with existing systems like Facebook and Google. Unless there is a drastic improvement in user experience, chatbots would not be able to drive conversations or delivery of services.