Human-AI interaction in safety-critical network infrastructures

Area of Science:

• Computer Science
• Decision-Making Sciences
• Psychology
• Industrial Practices

Background:

• Artificial Intelligence (AI) is revolutionizing society and critical infrastructures.
• AI integration enhances operational flexibility and resilience against hazards.
• The rise of AI necessitates a new paradigm for human-AI interaction.

Purpose of the Study:

• To outline challenges and requirements for effective human-AI interaction.
• To explore interdisciplinary approaches merging computer science, decision-making, psychology, and industry.
• To analyze emblematic safety-critical scenarios in energy and mobility.

Main Methods:

• Interdisciplinary review of human-AI interaction.
• Focus on safety-critical systems in power grids, railway networks, and air traffic management.
• Analysis of AI's role in decision-making and human collaboration.

Main Results:

• Identified critical challenges in defining roles for humans and AI systems.
• Highlighted the need for a well-defined ecosystem for healthy human-AI collaboration.
• Emphasized the importance of an interdisciplinary approach for effective integration.

Conclusions:

• Effective human-AI interaction is crucial for leveraging AI in safety-critical domains.
• Further research is needed to establish clear frameworks for human-AI collaboration.
• Successful integration requires addressing technical, psychological, and industrial aspects.