On 3 March the AI Observatory celebrates its first year in existence. We look back on an eventful first year with many research projects, numerous dialogues - and look forward to the upcoming projects.
On 3 March 2020, Federal Minister of Labour and Social Affairs Hubertus Heil opened the AI Observatory in front of 300 guests in the Café Moskau in Berlin. One year after the opening, we can look back on an eventful and successful first year of the AI Observatory. Above all, we want to thank everyone who has engaged with us – at our digital events or in the projects supported by the AI Observatory as well as to all those who contribute ideas and help us to close gaps in the field of AI research.
Indicators developed for the areas of work, skills and society
One of the main focus areas of the AI Observatory during 2020 was on the development of indicators, or key figures, to gain a better understanding of how AI affects the workplace and society. In this context, we are concentrating together with our partners on a knowledge gap because previous research into indicators has focused mainly on the field of economics.
Building on the insights from a virtual workshop held in mid- May 2020 with more than 40 stakeholders from politics, business, science and society, the AI Observatory developed ten indicators. The indicators describe the consequences of AI in the three areas of work, skills and society. Thanks to these key figures, we now know, among other things, that the majority of employees in Germany does not expect to lose their job because of AI. We know that machine learning expertise is the most important AI skill. We also know that the population in Germany is clearly against the use of AI in courts of law and in political decision-making. You can find more detailed information about the individual indicators in the new web area dedicated to the topic of AI indicators.
These insights not only provide a better understanding of AI in the workplace but also offer the basis for human-centric AI policies. In short: a key milestone in the first year of the AI Observatory.
Research project “ExamAI – AI Testing & Auditing” started
Audit and test procedures are important modules for using AI in a manner that is oriented towards the common good. Various scientific institutes have been researching these procedures since May 2020 as part of the project “ExamAI – AI Testing & Auditing”, which was initiated by the AI Observatory.
The researchers are investigating which audit and test procedures can be used to ensure security, transparency, equal opportunities and data protection in human-machine cooperation during industrial production and when AI is used for personnel and talent management. As a first step, the scientists developed application scenarios for this purpose. In the second step, they are now developing audit and test methods for the application of AI in both scenarios. Recommended actions will then be formulated based on the conclusions reached.
Dialogue held with experts; political decisions scientifically supported
Whether it is defining indicators or conducting research projects – the AI Observatory is constantly engaging with experts, including with the expert dialogue “HTI – Working with AI”. This initiative was launched in August 2020 and offers a platform for sharing technical and hands-on information. “HTI” stands for “Human-Technology Interaction” and describes the focus of the dialogue series. Here, experts discuss how the demand for human-centric AI use in operational practice can be implemented. The leading question: how can humans and machines become colleagues in practice and how do we manage innovations? In two different workshops with different thematic focuses, the experts address this question. Two workshops have already taken place, and another four are due to be held by mid-May 2021.
Discussions and expert dialogues such as this one yield recommended actions that the AI Observatory can also share with political stakeholders to support evidence-based policy formulation.
Creating safety and trust based on common regulations
Last year, the project of the Policy Lab Digital Work and Society was closely involved in developing the German Government's position statement in response to the white paper on AI prepared by the EU Commission. One of the key demands of the 28-page paper: we need an EU-wide regulatory framework that not only establishes regulations for safe and trustworthy AI, but also ensures that they are implemented.
Connected at international level
The AI Observatory’s involvement in drafting a position statement in response to ideas put forward by the EU Commission shows that those who discuss and research AI must adopt a cross-border mindset. With this in mind, the AI Observatory systematically developed its network at international level – for instance, as part of the Global Partnership on Artificial Intelligence (GPAI), which began in June 2020.
The GPAI is a committee of international experts that is affiliated with the OECD. In this committee, 18 countries and the EU Commission bundles their AI expertise. Germany is represented by the AI Observatory in cooperation with the Federal Ministry for Economic Affairs and Energy along with numerous other experts. The GPAI has set itself the task of evaluating international developments in AI and, most importantly, developing global quality standards for human-centric AI. On this basis, the GPAI will formulate recommendations in the near future.
We look forward to the second year of the AI Observatory!
The understanding of artificial intelligence must evolve at the same pace as the technology itself. The same applies to the AI Observatory: its scientific standards, extensive international network and its policy of exchanging information will continue to develop in its second year. After all, there is still so much to discover.
The indicators are a good example: ten key figures are already defined, more will follow. That is because the impact of AI on the workplace cannot be described using just ten indicators. To close additional research gaps, the AI Observatory will also promote the “AI Ultra” research product.
starting in March 2021. Here, researchers from the Fraunhofer Institute for Industrial Engineering IAO and the University of Stuttgart will investigate how companies can successfully introduce AI and organise the associated transformation processes in their businesses. Until now, empirical data and case studies have been lacking in this area. The workshop series of the expert dialogue “HTI – Working with AI” will also be continued in the next months.
In its first year, the AI Observatory has revealed some of the dimensions of AI which Federal Minister of Labour and Social Affairs Hubertus Heil referred to during his opening speech on 3 March 2020 – although we are still at a very early stage. In the future, therefore, the AI Observatory will continue to investigate the effects of AI in the workplace and in society and highlight needs for action – driven by dialogue, with a scientific foundation and backed by a strong and growing network.