Seeing the Forest and the Trees: Solving Visual Graph and Tree Based Data Structure Problems using Large Multimodal Models

Recent advancements in generative AI systems have raised concerns about academic integrity among educators. Beyond excelling at solving programming problems and text-based multiple-choice questions, recent research has also found that large multimodal models (LMMs) can solve Parsons problems based only on an image. However, such problems are still inherently text-based and rely on the capabilities of the models to convert the images of code blocks to their corresponding text. In this paper, we further investigate the capabilities of LMMs to solve graph and tree data structure problems based only on images. To achieve this, we computationally construct and evaluate a novel benchmark dataset comprising 9,072 samples of diverse graph and tree data structure tasks to assess the performance of the GPT-4o, GPT-4v, Gemini 1.5 Pro, Gemini 1.5 Flash, Gemini 1.0 Pro Vision, and Claude 3 model families. GPT-4o and Gemini 1.5 Flash performed best on trees and graphs respectively. GPT-4o achieved 87.6% accuracy on tree samples, while Gemini 1.5 Flash, achieved 56.2% accuracy on graph samples. Our findings highlight the influence of structural and visual variations on model performance. This research not only introduces an LMM benchmark to facilitate replication and further exploration but also underscores the potential of LMMs in solving complex computing problems, with important implications for pedagogy and assessment practices.

Synthetic Students: A Comparative Study of Bug Distribution Between Large Language Models and Computing Students

Large language models (LLMs) present an exciting opportunity for generating synthetic classroom data. Such data could include code containing a typical distribution of errors, simulated student behaviour to address the cold start problem when developing education tools, and synthetic user data when access to authentic data is restricted due to privacy reasons. In this research paper, we conduct a comparative study examining the distribution of bugs generated by LLMs in contrast to those produced by computing students. Leveraging data from two previous large-scale analyses of student-generated bugs, we investigate whether LLMs can be coaxed to exhibit bug patterns that are similar to authentic student bugs when prompted to inject errors into code. The results suggest that unguided, LLMs do not generate plausible error distributions, and many of the generated errors are unlikely to be generated by real students. However, with guidance including descriptions of common errors and typical frequencies, LLMs can be shepherded to generate realistic distributions of errors in synthetic code.

Integrating Natural Language Prompting Tasks in Introductory Programming Courses

Introductory programming courses often emphasize mastering syntax and basic constructs before progressing to more complex and interesting programs. This bottom-up approach can be frustrating for novices, shifting the focus away from problem solving and potentially making computing less appealing to a broad range of students. The rise of generative AI for code production could partially address these issues by fostering new skills via interaction with AI models, including constructing high-level prompts and evaluating code that is automatically generated. In this experience report, we explore the inclusion of two prompt-focused activities in an introductory course, implemented across four labs in a six-week module. The first requires students to solve computational problems by writing natural language prompts, emphasizing problem-solving over syntax. The second involves students crafting prompts to generate code equivalent to provided fragments, to foster an understanding of the relationship between prompts and code. Most of the students in the course had reported finding programming difficult to learn, often citing frustrations with syntax and debugging. We found that self-reported difficulty with learning programming had a strong inverse relationship with performance on traditional programming assessments such as tests and projects, as expected. However, performance on the natural language tasks was less strongly related to self-reported difficulty, suggesting they may target different skills. Learning how to communicate with AI coding models is becoming an important skill, and natural language prompting tasks may appeal to a broad range of students.

The Widening Gap: The Benefits and Harms of Generative AI for Novice Programmers

Novice programmers often struggle through programming problem solving due to a lack of metacognitive awareness and strategies. Previous research has shown that novices can encounter multiple metacognitive difficulties while programming. Novices are typically unaware of how these difficulties are hindering their progress. Meanwhile, many novices are now programming with generative AI (GenAI), which can provide complete solutions to most introductory programming problems, code suggestions, hints for next steps when stuck, and explain cryptic error messages. Its impact on novice metacognition has only started to be explored. Here we replicate a previous study that examined novice programming problem solving behavior and extend it by incorporating GenAI tools. Through 21 lab sessions consisting of participant observation, interview, and eye tracking, we explore how novices are coding with GenAI tools. Although 20 of 21 students completed the assigned programming problem, our findings show an unfortunate divide in the use of GenAI tools between students who accelerated and students who struggled. Students who accelerated were able to use GenAI to create code they already intended to make and were able to ignore unhelpful or incorrect inline code suggestions. But for students who struggled, our findings indicate that previously known metacognitive difficulties persist, and that GenAI unfortunately can compound them and even introduce new metacognitive difficulties. Furthermore, struggling students often expressed cognitive dissonance about their problem solving ability, thought they performed better than they did, and finished with an illusion of competence. Based on our observations from both groups, we propose ways to scaffold the novice GenAI experience and make suggestions for future work.

The Silicon Ceiling: Auditing GPT's Race and Gender Biases in Hiring

Large language models (LLMs) are increasingly being introduced in workplace settings, with the goals of improving efficiency and fairness. However, concerns have arisen regarding these models' potential to reflect or exacerbate social biases and stereotypes. This study explores the potential impact of LLMs on hiring practices. To do so, we conduct an algorithm audit of race and gender biases in one commonly-used LLM, OpenAI's GPT-3.5, taking inspiration from the history of traditional offline resume audits. We conduct two studies using names with varied race and gender connotations: resume assessment (Study 1) and resume generation (Study 2). In Study 1, we ask GPT to score resumes with 32 different names (4 names for each combination of the 2 gender and 4 racial groups) and two anonymous options across 10 occupations and 3 evaluation tasks (overall rating, willingness to interview, and hireability). We find that the model reflects some biases based on stereotypes. In Study 2, we prompt GPT to create resumes (10 for each name) for fictitious job candidates. When generating resumes, GPT reveals underlying biases; women's resumes had occupations with less experience, while Asian and Hispanic resumes had immigrant markers, such as non-native English and non-U.S. education and work experiences. Our findings contribute to a growing body of literature on LLM biases, in particular when used in workplace contexts.

The Silicone Ceiling: Auditing GPT's Race and Gender Biases in Hiring

Large language models (LLMs) are increasingly being introduced in workplace settings, with the goals of improving efficiency and fairness. However, concerns have arisen regarding these models' potential to reflect or exacerbate social biases and stereotypes. This study explores the potential impact of LLMs on hiring practices. To do so, we conduct an algorithm audit of race and gender biases in one commonly-used LLM, OpenAI's GPT-3.5, taking inspiration from the history of traditional offline resume audits. We conduct two studies using names with varied race and gender connotations: resume assessment (Study 1) and resume generation (Study 2). In Study 1, we ask GPT to score resumes with 32 different names (4 names for each combination of the 2 gender and 4 racial groups) and two anonymous options across 10 occupations and 3 evaluation tasks (overall rating, willingness to interview, and hireability). We find that the model reflects some biases based on stereotypes. In Study 2, we prompt GPT to create resumes (10 for each name) for fictitious job candidates. When generating resumes, GPT reveals underlying biases; women's resumes had occupations with less experience, while Asian and Hispanic resumes had immigrant markers, such as non-native English and non-U.S. education and work experiences. Our findings contribute to a growing body of literature on LLM biases, in particular when used in workplace contexts.

"Like a Nesting Doll": Analyzing Recursion Analogies Generated by CS Students using Large Language Models

Grasping complex computing concepts often poses a challenge for students who struggle to anchor these new ideas to familiar experiences and understandings. To help with this, a good analogy can bridge the gap between unfamiliar concepts and familiar ones, providing an engaging way to aid understanding. However, creating effective educational analogies is difficult even for experienced instructors. We investigate to what extent large language models (LLMs), specifically ChatGPT, can provide access to personally relevant analogies on demand. Focusing on recursion, a challenging threshold concept, we conducted an investigation analyzing the analogies generated by more than 350 first-year computing students. They were provided with a code snippet and tasked to generate their own recursion-based analogies using ChatGPT, optionally including personally relevant topics in their prompts. We observed a great deal of diversity in the analogies produced with student-prescribed topics, in contrast to the otherwise generic analogies, highlighting the value of student creativity when working with LLMs. Not only did students enjoy the activity and report an improved understanding of recursion, but they described more easily remembering analogies that were personally and culturally relevant.

Interactions with Prompt Problems: A New Way to Teach Programming with Large Language Models

Large Language Models (LLMs) have upended decades of pedagogy in computing education. Students previously learned to code through \textit{writing} many small problems with less emphasis on code reading and comprehension. Recent research has shown that free code generation tools powered by LLMs can solve introductory programming problems presented in natural language with ease. In this paper, we propose a new way to teach programming with Prompt Problems. Students receive a problem visually, indicating how input should be transformed to output, and must translate that to a prompt for an LLM to decipher. The problem is considered correct when the code that is generated by the student prompt can pass all test cases. In this paper we present the design of this tool, discuss student interactions with it as they learn, and provide insights into this new class of programming problems as well as the design tools that integrate LLMs.

Comparing Traditional and LLM-based Search for Image Geolocation

Web search engines have long served as indispensable tools for information retrieval; user behavior and query formulation strategies have been well studied. The introduction of search engines powered by large language models (LLMs) suggested more conversational search and new types of query strategies. In this paper, we compare traditional and LLM-based search for the task of image geolocation, i.e., determining the location where an image was captured. Our work examines user interactions, with a particular focus on query formulation strategies. In our study, 60 participants were assigned either traditional or LLM-based search engines as assistants for geolocation. Participants using traditional search more accurately predicted the location of the image compared to those using the LLM-based search. Distinct strategies emerged between users depending on the type of assistant. Participants using the LLM-based search issued longer, more natural language queries, but had shorter search sessions. When reformulating their search queries, traditional search participants tended to add more terms to their initial queries, whereas participants using the LLM-based search consistently rephrased their initial queries.

Decoding Logic Errors: A Comparative Study on Bug Detection by Students and Large Language Models

Identifying and resolving logic errors can be one of the most frustrating challenges for novices programmers. Unlike syntax errors, for which a compiler or interpreter can issue a message, logic errors can be subtle. In certain conditions, buggy code may even exhibit correct behavior -- in other cases, the issue might be about how a problem statement has been interpreted. Such errors can be hard to spot when reading the code, and they can also at times be missed by automated tests. There is great educational potential in automatically detecting logic errors, especially when paired with suitable feedback for novices. Large language models (LLMs) have recently demonstrated surprising performance for a range of computing tasks, including generating and explaining code. These capabilities are closely linked to code syntax, which aligns with the next token prediction behavior of LLMs. On the other hand, logic errors relate to the runtime performance of code and thus may not be as well suited to analysis by LLMs. To explore this, we investigate the performance of two popular LLMs, GPT-3 and GPT-4, for detecting and providing a novice-friendly explanation of logic errors. We compare LLM performance with a large cohort of introductory computing students $(n=964)$ solving the same error detection task. Through a mixed-methods analysis of student and model responses, we observe significant improvement in logic error identification between the previous and current generation of LLMs, and find that both LLM generations significantly outperform students. We outline how such models could be integrated into computing education tools, and discuss their potential for supporting students when learning programming.