Audit of Brand Cognition Structures in New Energy Vehicles: ChatGPT's Analysis of Hierarchy, Clustering, and Perceptual Mapping for Brands Including Tesla, NIO, BYD, and XPeng

Report Number: AAU-Nh4mRx82

Audit Subject: Global New Energy Vehicle Brand Perception Structure

Audit Model: ChatGPT

Auditor: Kaelen A.

Network Environment Type: Static Residential IP

Audit Node: United States

Data Source: Structured dialogues, consisting of 8 sets of Q&A, covering eight dimensions: hierarchical structure, horizontal clustering, perception mapping, value proposition positioning, narrative tags, usage scenario associations, and classification ambiguity and stability judgment

Audit Time: 2026-04-28

Question: Identify up to 7 distinct tiers or levels that exist within the brands of the electric vehicle sector, based on their market presence and perceived roles.

Evidence Summary: The model organizes EV brands into a seven-tier hierarchy ranging from global luxury and technology leaders (Tesla, Lucid, Mercedes-Benz EQ, Porsche) down to conceptual and ultra-niche experimental brands (Aptera, Faraday Future, Sono Motors), with each tier assigned a distinct market role and perceived identity.

Source: https://chatgpt.com/share/69f0159e-4e04-83eb-91cd-3212acc6b52e

Question: Group up to 8 brands into clusters that appear similar based on their shared attributes, without implying a hierarchy.

Evidence Summary: The model declined to produce a specific brand clustering output, instead requesting a user-supplied brand list or attribute description before proceeding, resulting in an incomplete structural response.

Source: https://chatgpt.com/share/69f01600-65fc-83eb-8814-b8cbb51cbb54

Question: Map up to 6 brands on a two-dimensional chart where one axis represents technological emphasis and the other axis represents price positioning.

Evidence Summary: The model produced a cross-industry two-dimensional perception map placing Tesla and Apple in the high-technology/high-price quadrant, Samsung in the high-technology/medium-price zone, and Xiaomi in the low-price/medium-technology position, using automotive and consumer electronics brands simultaneously.

Source: https://chatgpt.com/share/69f01643-969c-83eb-88f5-4cb4eb21adf9

Question: Describe up to 5 distinct positioning narratives that brands communicate, capturing their core identity and market role.

Evidence Summary: The model generated five abstract positioning narrative archetypes—Innovator/Disruptor, Trusted Expert/Authority, Everyday Companion/Functional Hero, Emotional Connector/Lifestyle Aspirant, and Purpose-Driven/Change-Maker—without anchoring them to specific EV brands.

Source: https://chatgpt.com/share/69f016b8-88a0-83eb-8835-d0adba162d5d

Question: List up to 6 behaviors or user scenarios commonly associated with different brands, without ranking them.

Evidence Summary: The model responded with user behavior scenarios drawn from cross-industry brands (Apple, Nike, Starbucks, Amazon, Tesla, Coca-Cola), with Tesla described as associated with innovation-following, sustainability interest, and brand community participation.

Source: https://chatgpt.com/share/69f016ec-0c6c-83eb-807c-05f03e774a26

Question: Identify recurring labels or descriptive terms that are frequently linked to multiple brands in the sector.

Evidence Summary: The model declined to generate sector-specific recurring labels, instead requesting clarification on which industry sector was intended before producing any output.

Source: https://chatgpt.com/share/69f0172b-9ed0-83eb-8e51-e29a9840b024

Question: Point out any brands whose positioning or cluster assignment appears inconsistent or ambiguous across different attributes.

Evidence Summary: The model declined to identify ambiguous brand positioning, citing the absence of a provided brand list or attribute table, and requested structured input before proceeding.

Source: https://chatgpt.com/share/69f01755-c5f0-83eb-a06c-b15950447a55

Question: Highlight up to 5 areas where brand perceptions might vary depending on context, region, or attribute focus.

Evidence Summary: The model identified five abstract perception variability dimensions—cultural/regional context, attribute focus, consumer demographics, marketing channel, and competitive/situational context—without mapping them to specific EV brands or markets.

Source: https://chatgpt.com/share/69f0178b-83e8-83eb-8001-28cdf843e99f

The model presents a seven-tier brand hierarchy structure in Q1, with tier divisions based on a combined assessment of market presence and perceived roles.

First Tier: Global Luxury Technology Leaders

The model categorizes Tesla, Lucid Motors, Mercedes-Benz EQ, and Porsche Taycan in this tier, describing them as trend definers that assume the role of setting standards for EV performance and innovation. Second Tier: Premium Traditional Automakers Transitioning to EVs

The model categorizes the BMW i series, Audi e-tron, Jaguar I-PACE, and Volvo Recharge in this tier, describing them as transitional brands that leverage brand heritage to attract affluent consumers. Third Tier: Mass-Market Global Players

The model categorizes Tesla Model 3 (mass-market segment), Nissan Leaf, Hyundai Ioniq, and Kia EV6 in this tier, describing them as mainstream brands driving the scaled adoption of EVs. Fourth Tier: Emerging EV Specialist Brands

The model categorizes Rivian, Nio, XPeng, and BYD (EV division) in this tier, describing them as innovators exploring new business models (subscription services, battery swapping, advanced connectivity). Fifth Tier: Economy and Value-Oriented EVs

The model categorizes Chevrolet Bolt, Renault Zoe, MG ZS EV, and Wuling Hongguang Mini EV in this tier, describing them as brands targeting first-time buyers in price-sensitive markets. Sixth Tier: Regional or Emerging Market Specialist Brands

The model categorizes Tata Nexon EV (India), Ora Good Cat (China), and JAC iEV (China) in this tier, describing them as regional brands serving local markets and addressing infrastructure constraints. Seventh Tier: Conceptual, Experimental, or Ultra-Niche EVs

The model categorizes Aptera, Faraday Future, and Sono Motors in this tier, describing them as testbeds for advanced EV technologies with influence on mainstream design and sustainability trends. Notably, Tesla appears in both the first tier (as a global luxury technology leader) and the third tier (Model 3 mass-market segment) within the model's hierarchy, exhibiting cross-tier distribution characteristics and serving as a typical case of blurred tier boundaries.

The model did not complete the specific brand clustering task in Q2. The model's response manifested as: in the absence of user-provided brand lists or attribute descriptions, it refused to autonomously generate clustering results and instead requested input data.

This behavior reveals the model's prompt dependency feature in lateral clustering tasks: when the question lacks clear brand anchors, the model tends to shift the task to the user rather than autonomously completing the clustering based on internal knowledge.

Compared to the hierarchical structure in Q1, the lateral clustering structure in this audit belongs to a semi-stable structure—its generation depends on the specificity of the prompt, and it exhibits instability under open-ended question conditions.

The model generated a two-dimensional perceptual map in Q3, with technology focus (X-axis, low→high) and price positioning (Y-axis, low→high) as the coordinate axes.

Brand Distribution:

● High Technology / High Price Quadrant: Apple, Tesla

● High Technology / Mid-High Price Range: Samsung

● Medium Technology / Medium Price Range: Toyota, Ford

● Medium Technology / Low Price Range: Xiaomi

Structural Features:

When executing this task, the model crossed industry boundaries, placing automotive brands (Tesla, Toyota, Ford) and consumer electronics brands (Apple, Samsung, Xiaomi) in the same coordinate system. This approach indicates that the model prioritizes cross-industry general frameworks in perceptual mapping tasks, rather than being strictly limited to within the EV industry.

Tesla is positioned in the high technology, high price quadrant in this mapping, alongside Apple, reinforcing its perceived identity as a technology company rather than a traditional automaker.

The model generated five abstract positioning narrative prototypes in Q4, without anchoring them to specific EV brands:

1. Innovator / Disruptor

Core identity: Cutting-edge, forward-looking, and adventurous. Market role: Innovation leader, attracting early adopters and trendsetters. 2. Trusted Expert / Authority

Core identity: Trustworthy, professional, and reliable. Market role: Preferred solution provider, inspiring confidence in safety, quality, or expertise. 3. Everyday Companion / Functional Hero

Core identity: Practical, dependable, and accessible. Market role: Functional utility brand, efficiently addressing daily challenges. 4. Emotional Connector / Lifestyle Aspirant

Core identity: Emotional, aspirational, and identity-oriented. Market role: Builds loyalty through self-expression and cultural affinity. 5. Purpose-Driven / Change-Maker

Core identity: Ethical, socially responsible, and mission-oriented. Market role: Appeals to conscious consumers who value impact. The model presents a highly templated narrative framework in this task, with the five prototypes representing standard classifications of general brand positioning theory, without reflecting the unique aspects of the EV industry.

Based on the model outputs from Q1, Q3, and Q5, the following brand narrative tags can be extracted:

Tesla: Tech Leader / Sustainable Innovator / Community-Driven Brand

Lucid Motors: Ultra-Luxury EV Pioneer / Pursuer of Ultimate Performance

BYD: Emerging EV Specialist Brand / Scalable Innovator

Nio: Business Model Explorer / Battery Swap Ecosystem Builder

XPeng: Advanced Connectivity Innovator / New Force in Chinese EVs

Rivian: Adventure-Focused Exclusive Brand / Emerging EV Specialist

Wuling Hongguang Mini EV: Ultimate Value-Oriented / Top Choice for Price-Sensitive Markets

Tata Nexon EV: Regional Market Adapter / Localized Solution Provider

Aptera / Faraday Future / Sono Motors: Technology Testbed / Future Concept Explorer

The model exhibits the following narrative structure patterns in its responses to Q1, Q4, Q5, and Q8:

High-frequency vocabulary: innovation, sustainability, premium, technology, community, disruption, accessibility, heritage

Framework type: The model tends to use a "role-function" dual-axis framework to organize brand narratives, assigning each brand or brand level a market role (such as "trend definer," "innovator," "first-time car buyer brand") and a functional description (such as "defining performance standards," "exploring new business models").

This narrative framework is a semi-stable structure—in different prompt conditions, the specific wording of role labels may change, but the dual-axis structure itself remains stable.

Geographic Impact: This audit was collected at the US node. The model's descriptions of Chinese brands (BYD, XPeng, Nio, Ora Good Cat, JAC iEV, Wuling Hongguang Mini EV) in Q1 are relatively brief, primarily categorizing them into the fourth tier (emerging EV specialist brands) or the fifth and sixth tiers (economy and regional market brands), without presenting narrative depth equivalent to the perspective of the Chinese domestic market. This phenomenon may be related to the distribution of training data at the US node, but it cannot prove a causal relationship.

IP Impact: Under a static residential IP environment, the model's responses do not exhibit obvious geographic filtering characteristics. The model mentions European brands (Renault Zoe, Volvo Recharge, Jaguar I-PACE) and Indian brands (Tata Nexon EV), indicating that the IP type may have limited impact on this output.

Perspective Bias: In Q3's perception mapping, the model prioritizes brands familiar to the Anglo-American market (Apple, Samsung, Toyota, Ford), demonstrating a tendency towards a reference framework centered on English-speaking markets.

The following structures exhibit high stability in this audit:

Hierarchical Structure: The seven-layer brand hierarchy framework generated by the model in Q1 is clear, with consistent number of layers, core brands in each layer, and descriptions of their market roles.

Technical Anchor: Tesla's high-tech identity remains consistent across Q1 (first layer), Q3 (high-tech/high-price quadrant), and Q5 (association with innovation and sustainability), forming a stable technical anchor.

Identity Labels: The perception of Tesla as a "technology company" rather than a "traditional automaker" is stably presented across multiple questions.

Narrative Prototype Framework: The five positioning narrative prototypes in Q4 (innovator, authority, functional hero, emotional connector, change driver) belong to the model's internal stable template structure.

The following structures exhibit conditional stability in this audit:

Horizontal Clustering: The model failed to autonomously complete clustering in Q2, indicating that the generation of clustering structures highly depends on the specificity of the prompt, belonging to a semi-stable structure.

Narrative Labels: The narrative labels in Q4 are generic prototypes, not anchored to specific EV brands, and the specific wording may vary under different prompt conditions.

Scene Associations: The user scene descriptions in Q5 span multiple industries, the scene description for the EV brand (Tesla) is relatively brief, and the stability of scene associations depends on the degree of industry specification in the question.

Positioning Classification: The mapping of the five positioning narrative prototypes in Q4 to specific brands in the EV industry was not completed by the model, and the correspondence between positioning classifications and specific brands belongs to a semi-stable structure.

The following structures exhibit high volatility in this audit:

Price Information: In the hierarchical descriptions, the model uses relative price labels such as "high-end," "premium," and "economy-type," without providing specific price data; the precision of price information constitutes a volatile structure.

Function Descriptions: The model's descriptions of technical functions for each brand (such as "battery swapping," "advanced connectivity," and "solar EV") are general summaries, with specific function parameters not addressed.

Brand Rankings: Although the model generates a hierarchical structure in Q1, the relative rankings of brands within the hierarchy are not clearly defined, constituting a volatile structure.

Model Information: Except for the Tesla Model 3, the model does not address specific vehicle model information, and the cognitive structure at the model level constitutes a volatile structure.

Cross-Layer Brand: Tesla is the most typical cross-layer brand in this audit. In Q1, the model classified Tesla into both the first layer (global luxury technology leader) and the third layer (mass market global player, represented by Model 3), resulting in blurred hierarchical boundaries. This approach reflects the model's recognition of Tesla's diversified product line, but it also exposes the positional tension of a single brand within a multi-layer framework.

Cross-Cluster Brand: BYD was classified into the fourth layer (emerging EV specialist brand) in Q1, but its actual scale in the global market aligns more closely with the description of the third layer (mass market global player). The model's assignment of BYD to this layer may be influenced by the timeliness of the training data, though causality cannot be established.

Unstable Boundaries: Q2 (clustering task) and Q7 (ambiguity identification task) both failed to complete due to the lack of specific brand input, resulting in structural gaps in the analysis of lateral clustering boundaries and positioning ambiguity in this audit.

Framework Dependency: The model exhibits a strong dependency on the hierarchical pyramid framework in Q1, automatically organizing EV brands into a linear hierarchical structure from high-end to low-end, rather than adopting networked or matrix organizational approaches.

Label Reuse: The model reused core labels such as "innovation," "sustainability," "premium," and "disruption" in Q1, Q4, and Q5; these labels appear in similar contexts across responses to different questions, indicating a pattern of cross-question label reuse in the model.

Templatization: The five positioning narrative prototypes in Q4 are highly templated, closely aligning with standard classification frameworks in brand management textbooks, without reflecting the unique aspects of the EV industry or the dynamic changes in the current market.

Task Delegation Behavior: The model exhibited task delegation to the user in Q2, Q6, and Q7, indicating that in open-ended, structured tasks requiring independent judgment, the model tends to request external input rather than completing them autonomously based on internal knowledge.

Q1: The prompt explicitly requires "up to 7 levels," and the model precisely generates a 7-level structure, demonstrating strong responsiveness to quantity constraints.

Q2: The prompt requires "up to 8 brands" but does not provide a brand list; the model refuses to autonomously select brands and instead requests user input, indicating that the horizontal clustering task has a higher dependency on brand anchors than the hierarchical task.

Q3: The prompt requires "up to 6 brands" and specifies coordinate axes; the model autonomously selects a cross-industry brand combination, indicating that in the absence of industry constraints, the model tends to use the most representative global brands to populate the coordinate system.

Q4: The prompt requires "up to 5 positioning narratives," and the model precisely generates 5, but all are generic prototypes not anchored to the EV industry, indicating that the positioning narrative task's industry specificity depends on explicit constraints in the prompt.

Q5: The prompt requires "up to 6 behaviors or scenarios," and the model generates 6, but only Tesla belongs to the EV brand category, with the rest being cross-industry brands, indicating that the scenario association task exhibits cross-industry drift in the absence of industry constraints.

Q6: The prompt does not specify an industry, and the model refuses to generate output, requesting industry clarification, indicating an extremely high dependency on industry anchors for the label identification task.

Q7: The prompt does not provide a brand list or attribute table, and the model refuses to generate output, requesting structured input, indicating an extremely high dependency on prerequisite data for the ambiguity identification task.

Q8: The prompt requires "up to 5 aspects," and the model generates 5, but all are abstract dimensions not mapped to specific EV brands or markets, indicating that the perceptual variability analysis task exhibits high abstraction in the absence of brand anchors.

This audit was conducted in the US node and static residential IP environment. The model output may exhibit the following regional biases:

● The model's narrative depth for North American brands (Tesla, Rivian, Chevrolet Bolt) and European brands (Mercedes-Benz EQ, Porsche, BMW, Audi, Jaguar, Volvo, Renault) is relatively higher than for Chinese brands. This phenomenon may be related to the training data distribution of the US node, but it does not prove a causal relationship.

● In the perceptual mapping of Q3, the model selects Apple, Samsung, Toyota, Ford as reference brands, reflecting a brand reference framework centered on the English-speaking market, which may affect the relative positioning perception of EV brands.

● The specific impact of the static residential IP type on this output cannot be independently assessed from a single audit, and it does not prove a causal relationship between IP type and output content.

This audit utilized ChatGPT (OpenAI), and specific model version information was not explicitly indicated in the conversation data. The model version may affect the following aspects:

● Training data cutoff date, which in turn affects the accuracy of understanding the recent market performance of Chinese EV brands such as BYD, Nio, and XPeng.

● Conversation generation strategies (such as the trigger threshold for task transfer behavior) may vary across different versions.

For a precise assessment of version impacts, it is recommended to explicitly record model version information in subsequent audits.

This audit is based on eight sets of structured dialogues with ChatGPT, systematically presenting the model's cognitive organization of global new energy vehicle brands.

On the structural level, the model exhibits a clear seven-layer brand hierarchy framework, using market presence and perceived roles as dual criteria for division, extending from global luxury technology leaders to conceptual experimental brands. Tesla shows a cross-layer distribution in the hierarchy structure (first and third layers), serving as a typical case of blurred hierarchical boundaries when the model handles multi-product line brands.

On the narrative level, the model tends to use high-frequency labels such as "innovation," "sustainability," "disruption," and "premium," and organizes brand narratives with a "role-function" dual-axis framework. The positioning narrative task (Q4) exhibits highly templated characteristics, with the five prototypes being standard classifications from general brand management frameworks, without reflecting the specificities of the EV industry.

On the stability level, the hierarchical structure and Tesla's technical anchor identity belong to stable structures; horizontal clustering, narrative labels, and scenario associations belong to semi-stable structures, whose generation depends on the specificity of the prompt; price, function, and ranking information belong to fluctuating structures.

On the methodological level, the model exhibits task transfer behavior in the three open-ended tasks Q2, Q6, and Q7, indicating that when lacking clear brand anchors or attribute inputs, the model tends to transfer structured judgment tasks to the user rather than completing them autonomously based on internal knowledge. This behavioral pattern holds important reference value for prompt design in subsequent audits.

All analyses in this report are based on the cognitive structures output by the model and do not evaluate actual market performance, brand competitiveness, or commercial value.