Question 1 · Verbal Ability and Reading Comprehension
MCQ
Common Description:
The passage below is accompanied by four questions. Based on the passage, choose the best answer for each question.
Historical analysis AI systems claim neutrality by processing centuries of scientific literature through "unbiased" algorithms. Yet these systems reproduce historical exclusion: an AI trained on 19th-century physics papers systematically undervalues contributions from women and non-Western scientists because their work was rarely published in dominant journals. The designers frame this as a "data problem," but it reveals the data neutrality myth—the illusion that algorithms can transcend the biases embedded in their training corpora.
More insidiously, such AI performs historical revision by retroactively applying modern standards to past contexts. When an AI flags a 1920s paper for "lack of diversity" in its citation network, it ignores how colonial power structures limited access to publishing venues. This creates a double bind: historical work is criticized for reflecting its era's constraints, while modern work citing only "diverse" historical sources appears artificially progressive. True historical understanding requires acknowledging that every dataset is a product of its socio-technical moment—a recognition that technical-social dualism actively suppresses by separating "data" from its human context.
The phrase "double bind" in the second paragraph refers to the situation where:
- A.
AI systems simultaneously promote and undermine historical accuracy
- B.
Researchers face contradictory expectations regarding historical context
- C.
Modern citation practices inherently replicate past exclusion
- D.
Algorithmic neutrality claims mask intentional bias
Step-by-Step Solution
Key idea: This contextual inference question requires unpacking a metaphor through multi-concept synthesis (historical revision + technical-social dualism).
Step 1: Locate the phrase: "double bind" describes criticism of historical work for reflecting era constraints while modern work citing "diverse" sources appears artificially progressive.
Step 2: Identify the contradiction:
- Historical work: Criticized for lacking diversity (but constrained by era)
- Modern work: Praised for diverse citations (but artificially curated)
Step 3: Evaluate options:
- A: Incorrect—AI doesn't promote accuracy; it distorts it.
- B: Correct—researchers face impossible choice: accept historical constraints (criticized) or curate modern citations (artificial).
- C: Partially true but misses the bind (contradictory expectations).
- D: Describes the data neutrality myth, not the double bind.
Step 4: B captures the core contradiction in researcher expectations.
Answer: B
Question 2 · Verbal Ability and Reading Comprehension
MCQ
Common Description:
The passage below is accompanied by four questions. Based on the passage, choose the best answer for each question.
Contemporary historical analysis tools claim objectivity through algorithmic processing of centuries of scientific literature, yet they systematically undervalue contributions from non-Western scholars. An AI trained on predominantly European academic journals interprets pre-20th century Chinese astronomical observations as "primitive superstition" rather than sophisticated scientific practice. When such systems flag historical texts for "lack of methodological rigor," they apply modern Western scientific standards retroactively, ignoring how knowledge production was shaped by local epistemic traditions.
The system's designers frame this as a "data completeness issue," but it reveals the data neutrality myth—the illusion that algorithms can transcend the biases embedded in their training corpora. This creates a historical double bind: past scholarship is criticized for not meeting contemporary standards, while modern reinterpretations that incorporate non-Western perspectives are dismissed as "anachronistic." True historical understanding requires acknowledging that every dataset is a product of its socio-technical moment—a recognition that technical-social dualism actively suppresses by separating "data" from its cultural context.
Which of the following best describes the historical double bind created by AI historical analysis tools as presented in the passage?
- A.
Historical scholarship is criticized for lacking modern methods, while modern reinterpretations are dismissed as unscientific
- B.
Non-Western scholarship is undervalued in historical analysis but overrepresented in contemporary studies
- C.
Historical data is insufficient for AI training, yet expanding the dataset creates new technical challenges
- D.
Western scientific standards are applied universally, but non-Western methods are recognized as equally valid
Step-by-Step Solution
Key idea: This contextual inference question requires unpacking a metaphor through multi-concept synthesis.
Step 1: Locate the phrase "historical double bind" and its explanation: "past scholarship is criticized for not meeting contemporary standards, while modern reinterpretations that incorporate non-Western perspectives are dismissed as 'anachronistic.'"
Step 2: Identify the contradictory expectations:
- Historical work: Criticized for lacking modern methods
- Modern reinterpretations: Dismissed as anachronistic/unscientific
Step 3: Evaluate options:
- A: Correctly captures both elements of the double bind as stated in the passage.
- B: Contradicted—the passage states non-Western perspectives are dismissed, not overrepresented.
- C: Focuses on technical issues, not the conceptual double bind described.
- D: Opposite of the passage—Western standards are applied universally without recognizing alternatives.
Step 4: Only A accurately reflects the specific contradictory expectations described in the passage.
Answer: A
Question 3 · Verbal Ability and Reading Comprehension
MCQ
Common Description:
The passage below is accompanied by four questions. Based on the passage, choose the best answer for each question.
The integration of AI into scientific peer review promises efficiency but risks distorting the epistemic ecosystem. Traditional peer review functions as a social technology where reviewers perform "invisible labor"—mentoring authors, contextualizing findings within disciplinary cultures, and negotiating methodological standards. When AI tools automate manuscript screening, they optimize for surface metrics (citation counts, keyword density) while ignoring these socio-cognitive dimensions. This creates a dangerous feedback loop: journals using AI screening increasingly publish papers that "game" the algorithm through strategic keyword stuffing, inadvertently rewarding methodological conformity over conceptual innovation.
Crucially, this isn't merely a technical limitation. The AI's training data reflects historical publication biases—overrepresenting certain methodologies and underrepresenting interdisciplinary work—thus amplifying existing power structures. As emergent order theory predicts, small initial biases in the system cascade into large-scale homogenization. The most insidious effect is automation complacency: editors assume AI handles "objective" screening, neglecting their role in maintaining epistemic diversity. Consequently, the very mechanism designed to accelerate science becomes a brake on paradigm-shifting research. True reform requires recognizing peer review as a complex adaptive system, not a mechanical process to be optimized.
Which of the following, if true, would most fundamentally invert the relationship between AI screening and scientific innovation described in the passage?
- A.
AI tools are trained on datasets balanced for methodological diversity
- B.
Editors use AI outputs as one input among many human judgments
- C.
AI screening prioritizes papers with high methodological novelty scores
- D.
Studies show AI-screened journals have higher retraction rates
Step-by-Step Solution
Key idea: This conceptual inversion question requires identifying what would reverse the causal mechanism (AI → reduced innovation).
Step 1: Identify the core causal chain: AI screening → optimization for surface metrics → reward for conformity → suppression of innovation.
Step 2: Inversion requires flipping the outcome: AI screening must actively promote innovation rather than suppress it.
Step 3: Evaluate options:
- A: Balanced training data reduces bias but doesn't invert the relationship (conformity may still be rewarded).
- B: Human oversight mitigates but doesn't invert the core dynamic (AI still prioritizes surface metrics).
- C: Directly reverses the mechanism—AI now rewards novelty, making innovation advantageous.
- D: Confirms the negative outcome described, strengthening rather than inverting the relationship.
Step 4: Only C transforms AI from innovation-suppressor to innovation-advantage.
Answer: C
Question 4 · Verbal Ability and Reading Comprehension
MCQ
Common Description:
The passage below is accompanied by four questions. Based on the passage, choose the best answer for each question.
The integration of AI into scientific peer review promises efficiency but risks distorting the epistemic ecosystem. Traditional peer review functions as a social technology where reviewers perform "invisible labor"—mentoring authors, contextualizing findings within disciplinary cultures, and negotiating methodological standards. When AI tools automate manuscript screening, they optimize for surface metrics (citation counts, keyword density) while ignoring these socio-cognitive dimensions. This creates a dangerous feedback loop: journals using AI screening increasingly publish papers that "game" the algorithm through strategic keyword stuffing, inadvertently rewarding methodological conformity over conceptual innovation.
Crucially, this isn't merely a technical limitation. The AI's training data reflects historical publication biases—overrepresenting certain methodologies and underrepresenting interdisciplinary work—thus amplifying existing power structures. As emergent order theory predicts, small initial biases in the system cascade into large-scale homogenization. The most insidious effect is automation complacency: editors assume AI handles "objective" screening, neglecting their role in maintaining epistemic diversity. Consequently, the very mechanism designed to accelerate science becomes a brake on paradigm-shifting research. True reform requires recognizing peer review as a complex adaptive system, not a mechanical process to be optimized.
Which of the following, if observed in a scientific journal, would most strongly support the author's argument about emergent order in AI-assisted peer review?
- A.
A gradual increase in interdisciplinary research publications over time
- B.
A consistent decline in the diversity of methodological approaches used in published papers
- C.
An increase in the number of papers rejected for technical formatting issues
- D.
A steady rise in the average citation count of published papers
Step-by-Step Solution
Key idea: This question requires identifying evidence that supports the emergent order argument through causal chain synthesis.
Step 1: Identify the emergent order claim: Small initial biases cascade into large-scale homogenization.
Step 2: Determine what would support this: Evidence of increasing homogenization over time as a result of AI screening.
Step 3: Evaluate options:
- A: Contradicts the argument—interdisciplinary work is underrepresented in the system.
- B: Correctly shows the homogenization predicted by emergent order theory (methodological conformity).
- C: Addresses technical issues, not the socio-cognitive dimensions central to the argument.
- D: Could occur without homogenization (e.g., through genuine quality improvements).
Step 4: Only B demonstrates the large-scale homogenization that emerges from small initial biases.
Answer: B
Question 5 · Verbal Ability and Reading Comprehension
MCQ
Common Description:
The passage below is accompanied by four questions. Based on the passage, choose the best answer for each question.
Contemporary AI models of animal cognition often commit the anthropomorphic projection trap by assuming human-like reasoning processes in non-human species. When an AI analyzes octopus problem-solving behavior, it maps actions onto human cognitive frameworks—interpreting tool use as evidence of "planning" rather than embodied cognition shaped by evolutionary trade-offs. This misattribution stems from cognitive offloading: researchers outsource interpretive labor to AI systems trained on human-centric datasets, inadvertently reinforcing the assumption that human cognition is the universal benchmark.
The error becomes systemic when such AI tools generate "objective" behavioral metrics for conservation policy. A recent project used AI-scored "cognitive complexity" to prioritize species for protection, ranking octopuses below primates despite their ecological significance. This ignores how octopus cognition evolved for fluid environments—where distributed neural processing (not centralized planning) confers survival advantages. By offloading cognitive assessment to AI, conservationists mistake human-like traits for universal intelligence, potentially diverting resources from ecologically vital but "less complex" species. The solution requires recognizing cognition as context-dependent adaptation, not a linear hierarchy.
The author would most likely agree with which of the following statements about AI's role in animal cognition research?
- A.
AI tools should be abandoned in favor of traditional observation methods
- B.
AI's human-centric framing distorts understanding of non-human cognition
- C.
Octopus intelligence is superior to primate cognition in fluid environments
- D.
Cognitive offloading inevitably improves research efficiency
Step-by-Step Solution
Key idea: This author-agreement question requires synthesizing the critique across technological and biological domains.
Step 1: Identify the core argument: AI imposes human cognitive frameworks, misrepresenting non-human cognition.
Step 2: Evaluate options against this thesis:
- A: Too extreme—the author advocates reform, not abandonment (mentions "solution requires recognizing").
- B: Directly aligns with "anthropomorphic projection trap" and "mistake human-like traits for universal intelligence."
- C: Overreach—the passage compares adaptations, not declares superiority.
- D: Contradicted—the author shows offloading causes distortion, not improvement.
Step 3: B captures the nuanced critique: AI's framing distorts, but isn't inherently unusable with proper context.
Answer: B