A speech assistance platform »made in Germany«
SPEAKER project starts development of an AI platform for B2B dialogue systems
Erlangen/Sankt Augustin. Whether in the healthcare, finance or industrial sectors, many companies can expect significant future prospects from speech assistants that are based on artificial intelligence (AI) and also comply with European standards of data security and data sovereignty. Under the leadership of the Fraunhofer Institutes for Integrated Circuits IIS and for Intelligent Analysis and Information Systems IAIS, the consortium partners of the SPEAKER project are now starting to develop a German speech assistance platform. The purpose of the project, which is funded by the German Federal Ministry of Economics and Energy within the framework of its AI innovation competition, is to provide infrastructure, technology modules and standards for voice-controlled dialogue systems to be used in business-to-business applications. All partners met virtually last week for the project kickoff and officially launched their speech assistance platform – a platform, by the way, that is »made in Germany«.
The natural interaction with technology via speech is becoming increasingly important for many industries and sectors of the economy – in healthcare, for example, speech dialogue systems can make it easier for doctors and nurses to retrieve patient data in a hands-free manner, operate medical equipment without physical contact and document diagnoses via speech input. In the industrial sector, speech assistants bring significant benefits to the digital inspection and quality assurance of machines, vehicles or infrastructure. Also, service and administration processes can be made more efficient with dialogue-based assistants. However, in all applications, technology and data sovereignty must be optimally harmonized so that German companies in particular can use voice assistance systems in a targeted and secure manner.
A voice assistance platform for German industry
»With SPEAKER, we are creating a solution tailor-made for German industry«, says Dr. Joachim Köhler, department head at Fraunhofer IAIS. »The data sovereignty of personal and company-relevant information lies solely with the German companies and complies with European standards for data security. By integrating core technologies from the Fraunhofer Institutes IIS and IAIS as well as many renowned project partners, we aim to establish a nationally leading voice assistance platform for companies«.
Oliver Hellmuth, department head at Fraunhofer IIS, adds: »Especially in the B2B world, our platform will be able to adapt to an industry’s specific terminology, workflows and needs. This makes it possible for our partners from large-scale industry, medium-sized companies, start-ups and research to use AI methods in their immediate environment«.
In close cooperation with future users, the project partners are developing the first pilot applications, which are tested directly in practice. The plan is to transfer the prototypes straight into productive use after three years.
New opportunities for human-machine interaction
A German speech assistance platform will enable a new level of quality in human-machine communication that takes into account Germany’s legal standards regarding data protection. Many concrete application scenarios have already been defined.
A health logbook, for example, would allow patients to document their state of health comprehensively and as accessibly as possible through the use of natural language. The information contained in the logbook is made available to the doctors treating the patients. This is of particular value in the diagnosis and treatment of chronic diseases that have no clearly indicative symptoms, such as rheumatic conditions. AI-based solutions help with the evaluation and aggregation of the data, thus enabling doctors to quickly make an accurate diagnosis.
Smart voice assistants can also be beneficial in customer service. They are available 24/7 to answer questions and provide information on, say, tax-related topics for clients of tax consulting firms – just one of several potential use cases in the financial sector. Speech assistant solutions can help in other sectors as well by reducing the workload of technical and operational staff. Processes and interactions – for example, the creation of maintenance protocols – will become more intuitive and robust by using voice input to store essential information during inspection or repair processes. Even complex root cause analysis can be supported by AI technologies.
Ecosystem comprising industry, medium-sized businesses, start-ups and research ensures innovative capability
»The SPEAKER platform will be embedded in a comprehensive ecosystem consisting of large-scale industry, medium-sized companies, start-ups and research partners, thus ensuring a high level of innovation«, Hellmuth explains. The Fraunhofer Institutes IAIS and IIS are in charge of setting up the platform and ecosystem. They already have corresponding technologies and many years of experience in the field of speech assistance technologies, as well as platforms (e.g. AI4EU – European Artificial Intelligence On-Demand Platform) and global marketing strategies for speech and audio technologies (e.g. mp3).
At the project kickoff last week, more than 50 members of the SPEAKER consortium came together virtually, including the consortium leaders (Fraunhofer IAIS and Fraunhofer IIS), the funding agency (German Federal Ministry of Economics and Energy), the lead partner (German Aerospace Center), the accompanying research institution (Institute for Innovation and Technology [iit]), LoeschHundLiepold Kommunikations GmbH (LHLK), KI Bundesverband e.V. and of course the other consortium partners. The meeting was an opportunity for the partners to introduce the applications and technology components that they contribute to the project. It also represented the first time the functional scope of the future voice assistance platform, which is based on the partners’ previously identified use cases, was presented.
The project’s implementation phase officially began on April 1, 2020. In 2019, SPEAKER was selected as one of 16 outstanding concepts in Artificial Intelligence as a Driver for Economically Relevant Ecosystems, the innovation competition held by the German Federal Ministry of Economics and Energy.