Invited Speaker---Dr. Martin Kurze
Technology & Innovation, Product Management, Deutsche Telekom, Germany
Director of Research & Innovation, Telekom Innovation Laboratories, Germany
Dr. Martin Kurze has a strong background in computer science. He holds degrees in this subject and in chemistry, both from “Freie Universität Berlin” where he also received his PhD in 1998. His main areas of work are human-computer interaction, software engineering and innovation scouting as well as innovation management. He joined Deutsche Telekom’s r&d subsidiary in 1997 where he conducted numerous innovation projects and established research programs. Today he works for Deutsche Telekom’s Technology and Innovation board level and is responsible for advanced use cases using cloud and telco assets, e.g. robotics, AI for robots and human-robot interaction as well as automated manufacturing. He developed products and solutions such as mobile operating systems and security frameworks. His responsibilities also include research and promotion of privacy enhancing technologies. Most recent works on cloud based artificial intelligence for industry and service/home robots complement his earlier achievements.
Privacy Aware use of Personal Information for AI and Data Mining
Europe is the origin of the (infamous) GDPR (General Data Protection Regulation) which for many researchers and businesses seems to be an “innovation killer”, especially when dealing with Big Data and Machine Learning using PII (personally identifiable information). But Europe is at the same time funding projects like SPECIAL dealing with the challenge using an interdisciplinary approach.
Project SPECIAL explored needs and hurdles for monetizing information that telcos collect in order to (technically) deliver their services in a different (commercial) context. E.g. location information is necessary to connect mobile users/customers to the next sell tower. The same information could be used for other purposes too. However, GDPR requires explicit and informed consent for such a re-use of PII. Users often hesitate to give this consent because they do not know who does what with their data.
Using a linked data approach and privacy-oriented policy language (together with respective policy enforcement tools), project SPECIAL offers a tool (set) to give users full transparency and control over their data. This makes it much more likely for the users to give consent to sharing their PII: Now users not only “believe” or “trust” that their data is used only as agreed, they can check it (using a privacy dashboard). Also, they can withdraw their consent or change the policy attached to specific data sets.
The result is a world where privacy is respected in the way GDPR requests it. At the same time, AI and ML based tasks can be carried out using PII. Several use cases (“pilots”) have been implemented and tested by the consortium. Industry partners from the telco sector and from information brokers verified that the approach actually works and that users accept it. Certain constraints apply (of course).
The talk will give at least one detailed example and an outlook to some more use cases that really might change the way we treat our PII (or the PII of our customers). Finally, an outlook is given showing how the project results can be used in real-world industrial context. Additional research questions will be brought up as well.