The paper under analysis is “The flight to safety-critical AI: Lessons in Ai safety from the aviation industry,” written by Hunt. The article empirically examines how economic and political competitive factors have influenced the pace and nature of AI research and development in aviation. The author highlights that few scholars have attempted to analyze the empirical evidence of a race to the bottom in the case of AI. A race to the bottom is a competitive scenario in which a corporation strives to lower the competition’s pricing by abandoning quality requirements or worker safety (sometimes defying legislation) or reducing labor expenses.
The findings demonstrate that the absence of AI-based laws and regulations in the aviation sector has hampered the deployment of safety-critical AI. As a result, technical safety requirements for AI are still in the early phases of development, with standard-setting agencies focusing on less complicated use cases such as path planning, preventative analysis, and decision assistance (Hunt, n.d.). However, the author claims that there is some indication that the aircraft industry is aiming for dominance in AI safety.
The paper has a solid structure and coherent content, although the methodology has an issue. Namely, the author has provided the sources of evidence in the article but has not established or explained any criteria for choosing and verifying their credibility and reliability. As a result, the author has included interviews with industry experts but has not covered the limitations that ensue from working with qualitative data. As such, the experts’ opinions on AI safety in aircraft might be biased. Since the author believes that the study’s audience includes governments, regulators, businesses, and academics, it would be reasonable to designate the potential inadequacies in the analysis results. Thus, although the paper is logically potent and contains profound arguments, the basis of the work and evidence is dependent on authoritative sources, which are not addressed in the methodology and conclusions.
Reference
Hunt, W. (n.d.). The flight to safety-critical AI: Lessons in Ai safety from the aviation industry. UC Berkley.