In 2024, artificial intelligence is the top topic for millions of companies and employees worldwide. AI applications contribute to widespread optimization of daily processes and provide support with repetitive tasks. But where there is great potential, there is also a significant risk. Hanna Hedges, Data Scientist at the coeo Group, explains the technical challenges and ethical questions that AI raises in the field of data science and how AI can be further developed.
Editorial Team: Exciting trends and developments in AI are…
Hanna Hedges: The availability of many different models for the general public. ChatGPT and DALL-E are just the tip of the iceberg, although they’ve helped foster a better understanding of the possibilities and limitations of AI among the public and management levels.
Many available open-source applications allow us, but also our competitors, to use AI to optimize processes. This means it's now important to seize this opportunity and become a leader in integrating AI into our daily work.
Editorial Team: Technical challenges in working with AI and data are…
Hanna Hedges: Complex. AI models like neural networks are a black box for the user, and decision-making processes are difficult to understand. This is especially problematic when handling security-critical and sensitive data. Additionally, models must be integrated into existing, sometimes complicated, infrastructures, which requires deep technical understanding and expertise.
Finally, AI models need to be updated, monitored, revised, and adapted to new requirements and structures. Tackling all these tasks requires close collaboration between IT experts, data scientists, analysts, and engineers.
Editorial Team: Important ethical concerns when working with AI in data science are…
Hanna Hedges: We don’t live in a perfect world, and data can be biased too. If data is not thoroughly checked and cleaned, AI processes can amplify bias and discrimination. Through an ethical framework, we must ensure that factors such as ethnicity, geographical origin, and gender play no role in decision-making.
Another important aspect of working with AI is ensuring the privacy and data protection of our clients and debtors. Appropriate security measures must be implemented when processing personal and sensitive data with AI.
We are not the only ones facing these challenges. The EU AI Act is the world’s first comprehensive AI regulation from a regulatory authority. The law includes guidelines for AI system providers and developers, as well as users, which also includes coeo. This is a good start toward more transparency and less discrimination, but it’s essential that these guidelines are continuously developed and monitored as AI and its applications evolve.
Editorial Team: AI influences the work in the data field at coeo by…
Hanna Hedges: Allowing us to tap into new possibilities to increase productivity. Particularly in smart services, we see great potential to reduce repetitive and time-consuming manual work through the integration of GPT models. Combining AI-driven email communication with professional personal contact can optimize processes for both coeo and our clients and debtors. The result is faster response times, precise information, and competent service—qualities that set us apart and make us an attractive service provider.
Additionally, in the area of legal decisions, we see the inclusion of AI prediction models as a way to process and review more cases and better forecast success probabilities, thereby improving the profitability and efficiency of the legal process.
Editorial Team: I see room for AI development at coeo because…
Hanna Hedges: We are still dedicating too little time and resources to AI integration. The initial costs and effort are, of course, discouraging, but it is an issue we must address sooner or later to remain competitive.
Cover: © Hanna Hedges