As of this autumn, the Riigikogu (the Estonian parliament) is using a new system to prepare verbatim reports of its sittings. The solution, which is based on speech recognition technology, has great export potential and could prove to be a useful tool for other public and private organisations.
The system is called Hans and it uses the speech recognition technology created at Tallinn University of Technology (TalTech) to perform the time-consuming work of stenographers. Hans records the debates taking place in the Riigikogu as sound files, which the speech recognition software then transcribes. After this, editors revise the text, which is finally published on the Riigikogu website.
Next clients – from the police to journalism
Hans was created by Estonian IT-company Finestmedia. For almost 20 years it has been contributing to the success of Estonian e-governance and has played a part in the creation of a number of innovative solutions in the private sector. The company’s Head of Division Lauri Esko says that creating Hans was an exciting challenge. “We’ve created a solution that has much wider potential than just the use it’s being put to at the moment,” he said. “Systems utilising speech recognition can be very effective tools in the work of courts, prisons and the police, but also in health care, research and journalism, where writing things down and producing detailed reports of conversations is an unavoidable necessity,” Esko added that the solution had great export potential. “We’ve already introduced Hans to delegations from several countries who are interested in writing the success story of their e-governance,” he said.
Modern technology improving efficacy
According to Ahto Saks, the Administrative Director of the Chancellery of the Riigikogu, the development of the system has taken two years from idea to the introduction, and improvements are ongoing. “To start with, Hans will be producing verbatim records of Riigikogu sittings and Question Time,” he explained. “But there are plans to use the system in the work of committees in the future as well.”
Tanel Alumäe, the head of the Laboratory of Language Technology at Tallinn University of Technology, says the new system shows how modern technology can help make routine work more effective.
Free web-based speech recognition
“We’ve been actively working on speech recognition for 15 years, and Estonian systems are currently being used by radiologists, researchers, and journalists,” Alumäe said. “Web-based speech recognition is freely available to everyone on the tekstiks.ee site as well. Estonian might be a small language, but the technology based on it is at a really high level!” Globally, speech recognition is usually widely used with larger languages.