The correct answer is B. Lack of standards . In the CT-AI syllabus discussion of safety and AI, the listed characteristics that make it more difficult to ensure AI-based systems are safe include complexity , non- determinism , probabilistic nature , self-learning , lack of transparency, interpretability and explainability , and lack of robustness . Therefore, options A, C, and D are directly aligned with the syllabus list.
"Lack of standards" may be relevant to governance, regulation, or organizational maturity, but it is not one of the syllabus-listed intrinsic characteristics that makes AI-based systems difficult to use in safety-related applications. The syllabus separately discusses standards and regulations in Chapter 1, including ISO/IEC work, GDPR, automotive safety standards, and related regulatory contexts. However, this does not make
"lack of standards" the correct answer for this specific question. The question is targeting the safety-related AI characteristics in Chapter 2, not the broader standards landscape.
References/topics: CT-AI Syllabus Chapter 2, Section 2.8 "Safety and AI."
=========

The correct answer is A. Genetic algorithms . In the CT-AI syllabus, Chapter 1, Section 1.4 "AI Technologies" lists multiple technologies used to implement AI, including fuzzy logic, search algorithms, reasoning techniques, neural networks, Bayesian models, decision trees, random forest, linear regression, logistic regression, clustering algorithms, genetic algorithms , and support vector machines. The syllabus also states that AI-based systems typically implement one or more of these technologies.
Option B, autonomy , is not an AI implementation technology; it is a quality characteristic describing the ability of a system to work without human oversight or intervention for sustained periods. Option C, classification , is a supervised learning problem type in which inputs are assigned to predefined classes, not itself an implementation technology in the syllabus taxonomy. Option D, reinforcement learning , is a form of machine learning, but the option that directly appears in the AI technology list is genetic algorithms .
References/topics: CT-AI Syllabus Chapter 1, Section 1.4 "AI Technologies"; Chapter 3, Section 3.1
"Forms of ML."
=========

Given the context of the self-learning nature and the need for real-time adaptability, option A is least likely to describe the fall in performance because it deals with acceptance criteria rather than real-time performance issues.