Exploring the Use of AI in Government Services: Can Chatbots Provide Reliable Advice?

Governments around the world are increasingly turning to artificial intelligence (AI) to improve their services and provide advice to citizens. The use of chatbots powered by AI has been a topic of great interest, as it promises to automate and enhance the efficiency of public service interactions.

However, the reliability of AI-powered chatbots has been called into question, with concerns about accuracy and accountability. The emergence of generative AI in recent years has raised hopes for more human-like advisors that can provide sophisticated responses to a wide range of questions. But this has also come with the risk of errors and nonsensical answers, known as hallucinations.

In the UK, the Government Digital Service (GDS) tested a ChatGPT-based chatbot called GOV.UK Chat, only to find inaccuracies in some of its responses. Similarly, Portugal developed a chatbot to provide information on legal matters, funded by the European Union’s Recovery and Resilience Facility (RRF). While the chatbot performed well in answering basic questions, it struggled with more complex inquiries, highlighting the limitations of current AI technology.

Experts like Colin van Noordt and Sven Nyholm have cautioned against over-reliance on AI chatbots, emphasizing the importance of human accountability and the potential dangers of erroneous recommendations. Estonia, known for its advanced digital services, is taking a different approach by using Natural Language Processing (NLP) algorithms for its chatbots. While NLP may lack the conversational quality of Large Language Models (LLMs) like ChatGPT, it offers greater control and transparency in how the chatbot operates.

As governments continue to explore the use of AI in public services, finding the right balance between efficiency and reliability remains a key challenge. While AI-powered chatbots hold great promise for improving accessibility and responsiveness, it is essential to address concerns around accuracy, accountability, and ethical considerations. Collaborating with experts in AI ethics and continuing to iterate and improve chatbot systems will be crucial in ensuring that they provide valuable and trustworthy information to citizens.