The digitization of nearly all processes within companies is continuing to advance. Emerging technical concepts like artificial intelligence (AI) and machine learning (ML) present new possibilities while also posing challenges for companies and their workforces. How can we ensure that the implementation and operation of these systems remain manageable, expanding the capabilities of individuals and companies rather than limiting them?
Both companies and their workforces face challenges arising from new data-driven algorithms, systems, and business models. Consequently, companies are grappling with questions such as: Do we possess sufficient skills within our organization to truly comprehend and master these technologies? What dependencies do we establish when we entrust external platforms with the collection, analysis, and evaluation of our data, including machine operating data? Conversely, what steps must we take to assume control over as much of this data processing and analysis as possible?
Employees ponder over their ability to master these new technologies with their current skills and competencies. Furthermore, they question the value of these skills and competencies in this context. How can they feel confident in dealing with technology whose inner workings they do not fully understand?
There is no magical formula that can comprehensively address all these questions. However, numerous individual building blocks already exist, which can be combined using suitable solution patterns tailored to specific operational situations. The most crucial among these building blocks, so to speak, is a systemic perspective: the technical solutions, operational organizational forms and processes, and the competencies of individual employees must align and be optimized together for practical success.
What is important in this optimization process? Three key criteria emerge: transparency and explainability, both in technical systems and organizational structures. Achieving this is not straightforward with AI-based systems, given their inherent opacity, which extends even to their developers. Nevertheless, existing technical solutions can help restore explainability, for instance, approximating algorithmic behavior through flowcharts. The other two criteria are certainty of action, ensuring that interactions with these systems yield intended outcomes with a high probability, and freedom of action, enabling users to choose from multiple possible actions instead of being constrained by the system.
To ensure that the design solutions within a company meet these criteria, the Institute for Innovation and Technology (iit) in Berlin has developed a step-by-step procedure that assists companies in finding the appropriate solution. This procedure involves guiding questions and expert advice. Specifically, for complex technical systems, Annelie Pentenrieder and colleagues at the iit have devised a method that empowers users to generate concrete, visually represented ideas about human-technology interfaces.