In the interview, Hans-Peter Sailer, Machine Learning Reply, emphasizes that it is important “not to rely entirely on AI systems, but to question them again and again”
DFI CLub: Mr. Sailer, on November 15, 2018, the Federal Government adopted its “Artificial Intelligence Strategy”. What has happened since then? How did the topic “AI” develop in Germany year 2020?
Hans-Peter Sailer: The past year of science was all about artificial intelligence . Germany is finally becoming increasingly concerned with AI: Not only are new professorships and courses with a focus on “AI” being created, but also the funding measures and increasing number of start-ups In Germany, this attracts highly qualified specialists in the digital field. Berlin and Munich in particular have established themselves as AI locations. Established companies can also benefit from this development. Increased awareness and an improved attitude towards artificial intelligence form the basis for further digital transformation.
DFI CLub: According to a current PwC study, only six percent of companies in Germany use artificial intelligence or are currently implementing AI systems. What are possible braking factors?
Sailer: In our view, the lack of skills and the non-cross-company or even missing strategy for the implementation of AI projects are the two biggest braking factors. Added to this are the reservations about the new technology and the uncertainty of many regarding the security aspects of AI systems.
The focus should therefore be on a uniform plan for the implementation and use of AI, taking into account the organizational structure with a supportive change management process. We support our customers with these topics and offer, for example, the development of competencies in our own company through our own program for training data scientists.
DFI CLub: For which industries is AI generally best suited and why? Aside from chatbots and digital assistants, what are the concrete application scenarios?
Sailer: Above all, internal work processes can be supported by AI process automation and AI-controlled decision-making and the effort can be reduced. These include robotic process automation and natural language processing, which are widely used. In addition, companies that use machines and systems are already seeing visible benefits from AI, such as predictive analytics. By combining sensors, IoT-Platforms and AI-controlled analysis tools can not only monitor companies’ plants, but also predict malfunctions and failures. The industry affiliation does not matter.
DFI CLub: How can you recognize a provider’s true AI skills?
Sailer: A provider specializing in AI should be able to cover the possibility of an end-to-end solution, i.e. an interlocking concept of corporate strategy and AI technology. The specialist knowledge should extend over an extensive portfolio of technologies and infrastructures in order to find the optimal solution for the company. It is also important that providers go beyond the POC and not only have pilot projects.
DFI CLub: What are classic pitfalls when setting up AI projects in everyday business? What are the stumbling blocks in the implementation?
Sailer: The entire corporate context should be included in AI projects. A requirement analysis that is initially lacking can later lead to many problems. Stakeholders, such as end users or customers, should therefore not be neglected in an AI project. The added value of the AI project should be clearly communicated in order to avoid resistance or rejection. Experience has shown that use case workshops are helpful, in which a roadmap is created based on specific use cases and implemented with multidisciplinary teams.
Big data has also been a major challenge for companies for a long time, making the right selection of data increasingly difficult. If the basis is not correct, even the best algorithm cannot deliver accurate results. For the selection of the right model, there are AI platforms such as the solution from Datarobot or tools such as Thoughtspot, which provide independent evaluations and highly automate use case development.
DFI CLub: What role do big data play for AI systems?
Sailer: Big data plays a big role for artificial intelligence because it feeds on large amounts of data. An AI system can evolve due to the variety and variety of data. With existing data quality, it analyzes and learns more successfully with big data . An AI system can identify key patterns and trends that are otherwise misinterpreted or even overlooked.
DFI CLub: Keyword “colleague AI”: How do employees see the use of artificial intelligence? Are your concerns well founded? Where will humans have advantages over algorithms in the future?
Sailer: As with the general topic of digitization, employees are also afraid of being replaced by AI. Most of the time, AI relieves employees of their work autonomously and creates more time for high-value activities. If employees understand the purpose of AI systems, they see the technology as helpful support. So the goal should be to respond to employees and point out opportunities where AI can make work easier. In the future, workers will increasingly work hand in hand with autonomous co-bots and be supported by them. This gives employees time for new activities and the associated great potential for new projects and services.
DFI CLub: What are the risks of AI systems?
Sailer: The misuse of data, the creation and evaluation of content as well as pure decisions based on the supposed intelligence of the machines should be viewed critically. It is important not to rely entirely on AI systems, but to question them again and again. This also applies to the definition of AI tasks.
DFI CLub: Keyword “AI governance”: How loud is the call for legal regulation of AI? To what extent, in your opinion, should there be laws and regulations regulating AI (keywords “ethics”, “security” and “data protection”?
Sailer: Transparency should be particularly emphasized on the topic of “AI governance” and is constantly being challenged. Companies should hence the answers to the questions “Who?”, “What?”, “How?”, “When?”, “Where?” and “Why?” with regard to their data and their later use Improve data security through clearly required data ownership with defined responsibilities and create trust.
DFI CLub: How important are you to the AI observatory, which is currently being set up, to be opened in spring 2020 at the Federal Ministry of Labor (BMAS)?
Sailer: As far as I know, eight manageable positions are initially planned, with the aim of creating an AI market analysis that determines where and how AI is used in Germany. Proposals for possible legal regulations should then be derived. This initially appears to be a pragmatic approach with the potential to legally and ethically integrate AI technologies.
DFI CLub: Why is AI important these days and what should it be able to do in 2020?
Sailer: AI is not an indispensable component that we encounter in many different ways in everyday life and improve process steps. Digital (voice) assistants, recommendation engines for music or film streaming, but also vacuum cleaning robots have found their way into households. In medicine, AI is used in image recognition to identify and track diseases. Mobility is certainly a big issue for Germany. Autonomous driving, navigation or traffic light control are made possible or optimized by AI. In 2020, AI will certainly get closer to the goal of mastering more tasks in a system, including through an automatically generated AI architecture.
Author Profile
- Amram is a technical analyst and partner at DFI Club Research, a high-tech research and advisory firm .He has over 10 years of technical and business experience with leading high-tech companies including Huawei,Nokia,Ericsson on ICT, Semiconductor, Microelectronics Systems and embedded systems.Amram focuses on the business critical points where new technologies drive innovations.
