Major Milestones in AI Evolution AIE Practice Questions (Page 5)

Questions 21–25

1. 21expert · hard

   A tech company is building a team to develop a new AI product. They want to include researchers who have historically driven major breakthroughs. Which combination of expertise is most likely to lead to innovative AI development?

   • AA team that combines symbolic reasoning with neural approaches, drawing on the strengths of both historical traditions.
   • BA team that focuses solely on generative models, as they are the latest trend.
   • CA team focused on symbolic reasoning and rule-based systems, following the expert systems tradition.
   • DA team focused on deep learning and neural networks, following the tradition of Hinton and others.
   Reveal Solution Discussion Select an answer first
2. 22application · medium

   A financial compliance team wants to deploy an AI system that can explain its credit-approval decisions to regulators. The team lead recalls that early AI systems used hand-coded rules and were fully transparent, while modern deep learning models are often opaque. Which approach should the team prioritize to best meet the regulatory requirement for explainability?

   • ADeploy an expert system with a rule-based inference engine that mirrors the historical approach to decision-making.
   • BImplement a generative model that can produce natural-language explanations for each decision.
   • CSelect a deep learning model with a high accuracy score and document the training data thoroughly.
   • DUse a modern deep learning model and rely on post-hoc explanation tools like LIME or SHAP.
   Reveal Solution Discussion Select an answer first
4. 23expert · medium

   A museum curator is designing an exhibit on the history of AI. The exhibit will include a section on the evolution of conversational AI, from early chatbots to modern generative models. The curator wants to highlight the key difference between ELIZA and modern LLMs. Which statement best captures this difference for the exhibit?

   • AELIZA was a hardware system, while modern LLMs are software-only.
   • BELIZA was developed in the 1990s, while modern LLMs were developed in the 1960s.
   • CELIZA used simple pattern matching, while modern LLMs use deep learning to generate responses based on vast training data.
   • DELIZA could understand context, while modern LLMs cannot.
   Reveal Solution Discussion Select an answer first
5. 24application · easy

   A software architect is designing a system to automate customer support. They are considering using a modern large language model (LLM) that can generate human-like responses. The architect recalls that earlier AI systems, such as ELIZA, also attempted conversational interaction but were limited. What is the most significant difference between ELIZA and modern LLMs that the architect should consider?

   • AELIZA was a hardware system, while modern LLMs are software-only.
   • BELIZA was developed in the 1990s, while modern LLMs were developed in the 1960s.
   • CELIZA used rule-based pattern matching, while modern LLMs use deep learning to generate responses based on vast amounts of training data.
   • DELIZA could understand context, while modern LLMs cannot.
   Reveal Solution Discussion Select an answer first
6. 25application · easy

   An AI history museum is designing a timeline exhibit. They want to show how each major milestone built on earlier work. Which sequence correctly orders the milestones from earliest to latest?

   • ATuring test (1950) → Expert systems (1970s) → Deep learning breakthrough (2012) → Generative models (2020s)
   • BDeep learning (2012) → Turing test (1950) → Expert systems (1970s) → Generative models (2020s)
   • CTuring test (1950) → Deep learning (2012) → Expert systems (1970s) → Generative models (2020s)
   • DExpert systems (1970s) → Turing test (1950) → Generative models (2020s) → Deep learning (2012)
   Reveal Solution Discussion Select an answer first