I have been working with speech recognition for 20 years, in the past mainly for professional dictation. Speech recognition in the field of documentation works somewhat differently, because here speech is already ready for writing. Punctuation marks and paragraphs are also dictated, and the speech software is trained accordingly for optimal recognition.
With the further development of artificial intelligence, new possibilities have opened up in recent years, namely for the use of AI-based speech recognition in other areas as well. For journalists who work a lot with interviews or podcasts, for public administration or companies where many meetings and sessions are recorded. This is how the idea came about to develop a platform that can automatically transcribe audio recordings and recordings of normal spoken language into text.
As head of product management, I am responsible for both the project and the team around it. GoSpeech was a very extensive project that we developed partly online in cooperation with an international team. My area of responsibility is very extensive, the product development alone is very time-consuming. You need suppliers for special technologies. Which partners are suitable for this? How can we bring the product to market? Making contacts, drawing up business plans and numerous consultations are all part of the process.
Let's first take a look at what artificial intelligence can do today, for example also speech recognition from GoSpeech. This is about so-called pattern recognition, i.e. an audio signal is to be recognised in such a way that it is either converted into text or interpreted as a command. Depending on this, I receive the sentence "what will the weather be like tomorrow" (speech recognition) or I am shown the weather forecast (voice command).
Artificial intelligence will continue to develop in the field of speech over the next few years. There are already projects where artificial intelligence can be used to automatically generate new texts, and people are in the process of training artificial intelligence so that it can recognise the most important points of a text and create a summary from them. I.e. AI learns to interpret. And that's where the road will lead.
In 10, 20 years, AI will probably be able to interpret not only language, but also movements and emotions, and then we will move in the direction of the glass human being.
Of course, AI-based applications like Alexa, Siri or the navigation system in the car are very practical in everyday life, but on the other hand, such AI models also require incredible amounts of data that are collected and of course also used commercially. What will happen with this data?
At Grundig Business Systems, i.e. also at GoSpeech, absolute data security applies. The data remains on our servers, it is neither used for other purposes, nor passed on or evaluated. The same applies to our technology suppliers with whom we have collaborated for the development of GoSpeech.
One is the automated correction, the automatic editing of transcripts. Another area is the development of interfaces to enable other systems to use GoSpeech speech technology. Also exciting is the automatic indexing of media archives via speech recognition for later research or use as a source, which was previously only possible on the basis of sometimes very incomplete manually created protocols.
The further development of GoSpeech is strongly based on user requirements.
I actually try to switch off my computers sometimes and throw myself into real life - gardening, climbing, diving, my family. I have three children who keep me on my toes, literally. Just recently I built a new stable with my daughter for her horses.
Are you interested in our transcription software? Find out more about the advantages or take advantage of the opportunity to test GoSpeech for free right now!