The Dutch government is investing in innovation by organizing – amongst other things – Innovation Boosters. Recently Zoom Media and its newly founded sister Zoom Script attended one of these Innovation Boosters hosted by the Shared Service Center ICT (SSC-ICT). The SSC-ICT aims to link its knowledge of IT, and that of the private sector, with governmental institutions such as ministries. While sometimes the notion exists that certain innovations are something we might encounter in the future, the SSC-ICT tries to show that these innovations are already right in front of us.
One of those innovations, Speech Recognition, was on the agenda during the Innovation Booster Zoom Media and Zoom Script attended. The goal was to show in what ways a technology such as Speech Recognition can help make certain processes more efficient, less time consuming and how it can contribute to specific tools that already exist.
One of the most tedious jobs out there might very well be to transcribe interviews. It takes a lot of valuable time out of someone’s schedule which can be spend on much more important tasks. Moreover the time to take notes or to transcribe during an interview draws the attention away from the conversation. During the innovation booster Zoom Script explained how, by applying Speech Recognition, time can be saved during this process. Zoom Script enables individual users to upload their interviews using their own client dashboard. The content is then processed by the Zoom Media Speech Recognition engine and sent back almost instantaneously. Now the user has the option of editing the output received from the Speech Recognition engine or request a perfect transcript which will be handled by the Zoom Script team of language professionals. It’s always important to bear in mind that, with Speech Recognition, the results are never 100% accurate. However, the bulk of the work is done very quickly, leaving the user with not much more to do than applying some minor corrections thus saving a lot of valuable time.
Politicians love to debate and we love to watch them do so, but writing down every single word they say can be quite a hassle. Now since the government is obligated to make written reports of everything that has been said in parliament, one can imagine the strain this puts on governmental departments this task has been assigned to. During the Innovation Booster Arbor Media showed how Speech Recognition from Zoom Media can be used to create live closed captioning during debates. Speech Recognition from Zoom Media works in real-time, which means that every word can be converted into text thus subtitling the moment someone says it. This offers a great solution for those who are hard of hearing or hearing impaired to follow debates live as well. Additionally, using Speech Recognition on parliamentary debates offers the possibility of creating archives of text which are linked to the respective debates thanks to timestamps. This means anyone with access to such an archive – the citizens – can browse through it freely as if they were using a search engine. This way the decision-making process of governments can become more transparent and accessible.
Public safety is probably one of the most important tasks for any government. Pressure on safety authorities is therefore always a factor. At the same time however its difficult for these authorities to innovate since the interests are widespread and varied, options to innovate are widespread and varied as well and there is always the issue – thankfully – of privacy. So how does a niche innovation such as Speech Recognition fit in here? How can Speech Recognition offer a way for safety authorities such as the police to innovate? One of the possibilities Zoom Media discussed during the Innovation Booster is police hearings. All hearings conducted in the Netherlands are typed out manually, which means an enormous strain on the budget and a lot of time that could be spent on more important tasks, such as actual police work. Applying Speech Recognition would allow to convert all those hearings into text after which an interface such as that of Zoom Script can be used to correct the remaining minor errors. Privacy remains as an issue, however, Zoom Media Speech Recognition runs on Azure Cloud which meets all (inter)national privacy standards. Not only that, thanks to Azure Cloud, Zoom Media Speech Recognition can run on-premises, allowing the police in this example to use the technology in-house, keeping all data safe and secure.
We really enjoyed talking and learning about all the possibilities Speech Recognition has to offer the government and with it the public. Want to learn more? Visit our website or drop me an email at email@example.com.