Failures, liability issues and hacker attacks are unresolved challenges according to the survey
The Swiss energy and health sector is already using artificial intelligence and will expand corresponding applications in the coming years. This is the result of a survey conducted by the independent think tank Pour Demain with experts from almost 40 companies in the energy and healthcare sector. In addition to great potential, the respondents see significant challenges.
There is more than just ChatGPT: Two of the sectors critical to the well-being of the Swiss population, health and energy, have long been using a wide range of artificial intelligence (AI) applications - from prototypes for predicting energy availability to indispensable ones medical tools for diagnosing skin lesions.
Switzerland has so far lacked an overview of the use of AI in business and society. Therefore, as a pilot project, the think tank Pour Demain conducted a survey with experts from 38 companies from the energy and healthcare sectors.
The survey shows a growth trend: Four out of five respondents expect that AI will be used more extensively in these sectors in the next two years.
Potential in cancer detection and maintenance of power plants
Respondents expect many benefits, with particular emphasis on improved professional decision-making in clinical practice and predictive analytics for power plants and power grids. Several hospitals use AI for image recognition in radiology and dermatology, while energy suppliers use it, e.g. for production forecasts and maintenance of power plants.
Significant challenges and risks
However, the experts surveyed are concerned about significant challenges and risks that can arise in connection with AI systems: unexpected failures, cyberattacks, reproduced discrimination, liability issues or inexplicable recommendations.
Call for clarity of framework conditions
The need for clear AI frameworks and guidelines at economic and political level, as well as data protection and transparency are the most frequently mentioned with regard to pending regulatory measures. Doctors, for example, need clarity about liability issues when using AI systems that have a direct impact on patient health; Power grid operators rely on clear guidelines to establish trust when exchanging data with other actors.
Recommendations for industries and politics
Based on this survey, which was conducted for the first time, the think tank Pour Demain has developed the following recommendations.
Monitoring instead of flying blind: The Confederation and cantons need a better understanding of the use, opportunities and risks of AI in Switzerland. Like Diseases or Energy supply key figures are already being collected and published, a Switzerland-wide one is also needed with regard to AI monitoring. In this way, innovations can be promoted better and challenges can be overcome more quickly.
Review of existing legislation: It should be clarified whether current legislation is designed to deal with AI systems in the energy and healthcare sectors. The focus should be on powerful systems like ChatGPT. Policymakers can address concerns about explainability of AI decisions, reproduced discrimination and unpredictable system failures through guidelines, thus reducing uncertainty for organizations that want to use AI responsibly.
Set clear standards: Companies and industry associations should support ongoing standardization processes for trustworthy development, procurement and deployment of AI. The federal administration can accompany and promote these processes at the national and international level.
Education offensive: Companies and industry organizations should continue to expand training courses on AI. In addition, energy and health experts see an AI knowledge deficit among political decision-makers. Appropriate further training could contribute to responsible and forward-looking use of AI technologies at national level.
Study and Executive Summary
David Marti, Program Officer Artificial Intelligence Pour Demain
david.marti (at) pourdemain.ch
About Pour Demain
Pour Demain is a non-profit think tank that mediates between science, politics and society. We take up previously neglected topics such as artificial intelligence and develop concrete, science-based proposals - with a positive effect for Switzerland and beyond. www.pourdemain.ch