Michael Muller

IBM Research · Artificial Intelligence

Framing and Reframing are powerful co-creative strategies. We explore Reframing as a novel human-AI co-creative method using a conversational UI to a Large Language Model, presenting an actual session in the form of a theatrical script. We address themes of human-AI co-creativity, the role of the AI, and questions of AI anthropomorphism.

The locus of co-creativity in human-AI creative tasks has not been resolved. We explore a sub-part of this problem through the use of analogies for reframing a conceptual design task. In our scenarios, the human proposes an analogy, and then the human and a conversational UI to a Large Language Model (LLM) collab-oratively explore design features b...

This panel aims to generate conversation toward creating a more equitable CHI. In recognizing our community's hard work thus far, this panel seeks to engage panelists and participants with thought-provoking questions to garner and promote actionable items for the community. We intend to have an open dialogue on allyship, diversity, equity, and incl...

In recent years, the CHI community has seen significant growth in research on Human-Centered Responsible Artificial Intelligence. While different research communities may use different terminology to discuss similar topics, all of this work is ultimately aimed at developing AI that benefits humanity while being grounded in human rights and ethics,...

Large language models (LLMs) have recently been applied in software engineering to perform tasks such as translating code between programming languages, generating code from natural language, and autocompleting code as it is being written. When used within development tools, these systems typically treat each model invocation independently from all...

Fig. 1. Changes in human roles related to AI-based workflow systems. Alt-text: During the Research phase, the Data Scientist is likely to be most influential, followed by Designer and perhaps Configurator (customizer) During Development, Emphasis shifts to the Engineer and the Designer. During Distribution and Sales, the Configurator's concerns bec...

Generative AI technologies are growing in power, utility, and use. As generative technologies are being incorporated into mainstream applications, there is a need for guidance on how to design those applications to foster productive and safe use. Based on recent research on human-AI co-creation within the HCI and AI communities, we present a set of...

The Programmer's Assistant is an experimental prototype software development environment that integrates a chatbot with a code editor. Conversational capability was achieved by using an existing code-fluent Large Language Model and providing it with a prompt that establishes a conversational interaction pattern, a set of conventions, and a style of...

Conversational interfaces require two types of curation: data curation by data science workers and content curation by domain experts. Recent years have seen the possibilities for content curators to instruct conversational machines in the customer service domain (i.e., Machine Teaching). The activities of curating specialized data are time-consumi...

Geographically distributed teams often face challenges in coordination and collaboration, lowering their productivity. Understanding the relationship between team dispersion and productivity is critical for supporting such teams. Extensive prior research has studied these relations in lab settings or using qualitative measures. This paper extends p...

Computational notebooks allow data scientists to express their ideas through a combination of code and documentation. However, data scientists often pay attention only to the code, and neglect creating or updating their documentation during quick iterations. Inspired by human documentation practices learned from 80 highly-voted Kaggle notebooks, we...

HCI engages with data science through many topics and themes. Researchers have addressed biased dataset problems, arguing that bad data can cause innocent software to produce bad outcomes. But what if our software is not so innocent? What if the human decisions that shape our data-processing software, inadvertently contribute their own sources of b...

HCI engages with data science through many topics and themes. Researchers have addressed biased dataset problems, arguing that bad data can cause innocent software to produce bad outcomes. But what if our software is not so innocent? What if the human decisions that shape our data-processing software, inadvertently contribute their own sources of b...

Best practices for addressing the bias and inequality that may result from the automated collection, analysis, and distribution of large datasets. Human-centered data science is a new interdisciplinary field that draws from human-computer interaction, social science, statistics, and computational techniques. This book, written by founders of the f...

Generative machine learning models have recently been applied to source code, for use cases including translating code between programming languages, creating documentation from code, and auto-completing methods. Yet, state-of-the-art models often produce code that is erroneous or incomplete. In a controlled study with 32 software engineers, we exa...

What does it mean for a generative AI model to be explainable? The emergent discipline of explainable AI (XAI) has made great strides in helping people understand discriminative models. Less attention has been paid to generative models that produce artifacts, rather than decisions, as output. Meanwhile, generative AI (GenAI) technologies are maturi...

Human-Centered AI (HCAI) is an emerging discipline that aims to create AI systems that amplify [56, 55] and augment [58] human abilities and preserve human control in order to make AI partnerships more productive, enjoyable, and fair [25]. Our workshop aims to bring together researchers and practitioners from the NeurIPS and HCI communities and oth...

Translating source code from one programming language to another is a critical, time-consuming task in modernizing legacy applications and codebases. Recent work in this space has drawn inspiration from the software naturalness hypothesis by applying natural language processing techniques towards automating the code translation task. However, due t...

Geographically dispersed teams often face challenges in coordination and collaboration, lowering their productivity. Understanding the relationship between team dispersion and productivity is critical for supporting such teams. Extensive prior research has studied these relations in lab settings or using qualitative measures. This paper extends pri...

Explainability of AI systems is critical for users to take informed actions and hold systems accountable. While "opening the opaque box" is important, understanding who opens the box can govern if the Human-AI interaction is effective. In this paper, we conduct a mixed-methods study of how two different groups of whos-people with and without a back...

Employee engagement (EE) has been shown to have important implications for the success of organizations. Most researchers have discussed employee engagement in terms of factors in a top-down, hierarchical model of the organization. However, there may also be contributing factors from an employee‟s intra-organizational social network. (IOSN). In thi...

We describe a first experiment in enterprise crowdfunding –i.e., employees allocating money for employee-initiated proposals in an Intranet site, including a trial of this system with 511 employees in a large multinational company. Major outcomes include employee proposals addressing diverse individual and organizational needs, high participation r...

Social media has led to a data explosion and has begun to play an ever increasing role as a valuable source of information and a mechanism for information discovery. The wealth of data highlights the need for methods to filter and sort information in order to allow users to discover useful information. Most traditional solutions focus on the user,...

Social search has been claimed to improve content discovery by allowing users to draw on their social network to find relevant content. Thus social network information, complemented with metadata, can enhance the search for new information. We examine the relative contribution of social network information and file metadata in predicting downloads...

Explainability of AI systems is critical for users to take informed actions and hold systems accountable. While "opening the opaque box" is important, understanding who opens the box can govern if the Human-AI interaction is effective. In this paper, we conduct a mixed-methods study of how two different groups of whos--people with and without a bac...

Ground-truth labeling is an important activity in machine learning. Many studies have examined how crowdworkers apply labels to records in machine learning datasets. However, there have been few studies that have examined the work of domain experts when their knowledge and expertise are needed to apply labels. We provide a grounded account of the w...

As AI-powered systems increasingly mediate consequential decision-making, their explainability is critical for end-users to take informed and accountable actions. Explanations in human-human interactions are socially-situated. AI systems are often socio-organizationally embedded. However, Explainable AI (XAI) approaches have been predominantly algo...

The development of AI applications is a multidisciplinary effort, involving multiple roles collaborating with the AI developers, an umbrella term we use to include data scientists and other AI-adjacent roles on the same team. During these collaborations, there is a knowledge mismatch between AI developers, who are skilled in data science, and exter...

Human labeling of training data is often a time-consuming, expensive part of machine learning. In this paper, we study "batch labeling", an AI-assisted UX paradigm, that aids data labelers by allowing a single labeling action to apply to multiple records. We ran a large scale study on Mechanical Turk with 156 participants to investigate labeler-AI-...

Generative models have become adept at producing artifacts such as images, videos, and prose at human-like levels of proficiency. New generative techniques, such as unsupervised neural machine translation (NMT), have recently been applied to the task of generating source code, translating it from one programming language to another. The artifacts p...

Labeling data is an important step in the supervised machine learning lifecycle. It is a laborious human activity comprised of repeated decision making: the human labeler decides which of several potential labels to apply to each example. Prior work has shown that providing AI assistance can improve the accuracy of binary decision tasks. However, t...

Computational notebooks allow data scientists to express their ideas through a combination of code and documentation. However, data scientists often pay attention only to the code, and neglect creating or updating their documentation during quick iterations, which leads to challenges in sharing their notebooks with others and future selves. Inspire...

The development of AI applications is a multidisciplinary effort, involving multiple roles collaborating with the AI developers, an umbrella term we use to include data scientists and other AI-adjacent roles on the same team. During these collaborations, there is a knowledge mismatch between AI developers, who are skilled in data science, and exter...

As AI-powered systems increasingly mediate consequential decision-making, their explainability is critical for end-users to take informed and accountable actions. Explanations in human-human interactions are socially-situated. AI systems are often socio-organizationally embedded. However, Explainable AI (XAI) approaches have been predominantly algo...

Data science and machine learning (DS/ML) are at the heart of the recent advancements of many Artificial Intelligence (AI) applications. There is an active research thread in AI, \autoai, that aims to develop systems for automating end-to-end the DS/ML Lifecycle. However, do DS and ML workers really want to automate their DS/ML workflow? To answer...

Recently, the automated translation of source code from one programming language to another by using automatic approaches inspired by Neural Machine Translation (NMT) methods for natural languages has come under study. However, such approaches suffer from the same problem as previous NMT approaches on natural languages, viz. the lack of an ability...

Recent progress in machine learning has given rise to a plethora of tools and applications that rely on conversational interactions, from chatbots, speech-controlled devices to robots and virtual agents. Conversational interfaces are becoming widely accepted for utility tools, where a common function is to serve users’ information needs. Albeit wit...

Today, the prominence of data science within organizations has given rise to teams of data science workers collaborating on extracting insights from data, as opposed to individual data scientists working alone. However, we still lack a deep understanding of how data science workers collaborate in practice. In this work, we conducted an online surve...

Generative AI is a class of machine learning technology that learns to generate new data from training data. While deep fakes and media-and art-related generative AI breakthroughs have recently caught people's attention and imagination, the overall area is in its infancy for business use. Further, little is known about generative AI's potential for...

Today, the prominence of data science within organizations has given rise to teams of data science workers collaborating on extracting insights from data, as opposed to individual data scientists working alone. However, we still lack a deep understanding of how data science workers collaborate in practice. In this work, we conducted an online surve...

We explore trust in a relatively new area of data science: Automated Machine Learning (AutoML). In AutoML, AI methods are used to generate and optimize machine learning models by automatically engineering features, selecting models, and optimizing hyperparameters. In this paper, we seek to understand what kinds of information influence data scienti...

Social media platforms and social network sites generate a multitude of digital trace behavioral data, the scale of which often necessitates the use of computational data science methods. On the other hand, the socio-behavioral and often relational nature of the social media data requires the attention to context of user activity traditionally asso...

Artificial Intelligence (AI) can now automate the algorithm selection, feature engineering, and hyperparameter tuning steps in a machine learning workflow. Commonly known as AutoML or AutoAI, these technologies aim to relieve data scientists from the tedious manual work. However, today's AutoAI systems often present only limited to no information a...

Two general routes have been followed to develop artificial agents that are sensitive to human values---a top-down approach to encode values into the agents, and a bottom-up approach to learn from human actions, whether from real-world interactions or stories. Although both approaches have made exciting scientific progress, they may face challenges...

In society today, people experiencing disability can face discrimination. As artificial intelligence solutions take on increasingly important roles in decision-making and interaction, they have the potential to impact fair treatment of people with disabilities in society both positively and negatively. We describe some of the opportunities and risk...

In recent years there has been an increasing trend in which data scientists and domain experts work together to tackle complex scientific questions. However, such collaborations often face challenges. In this paper, we aim to decipher this collaboration complexity through a semi-structured interview study with 22 interviewees from teams of bio-medi...

This paper reflects on the expectations of museum guides regarding companion AI-powered robots in a science museum space. We employed Design Fiction as a technique to explore machine teaching of future technologies in public spaces. The fiction is illustrated by an open-ended “imaginary abstract” which showcases the dilemma of buying AI robots to w...

Qualitative methods have long been an important component of CSCW research. However, it can be challenging to make qualitative work legible to a broader set of researchers, which is critical as mixed methods research becomes more common. Moreover, the shift towards larger scales of data and increasing calls for open data and more transparency pose...

The rapid advancement of artificial intelligence (AI) is changing our lives in many ways. One application domain is data science. New techniques in automating the creation of AI, known as AutoAI or AutoML, aim to automate the work practices of data scientists. AutoAI systems are capable of autonomously ingesting and pre-processing data, engineering...

This paper reflects on the expectations of museum guides regarding companion AI-powered robots in a science museum space. We employed Design Fiction as a technique to explore machine teaching of future technologies in public spaces. The fiction is illustrated by an open-ended "imaginary abstract" which showcases the dilemma of buying AI robots to w...

In recent years there has been an increasing trend in which data scientists and domain experts work together to tackle complex scientific questions. However, such collaborations often face challenges. In this paper, we aim to decipher this collaboration complexity through a semi-structured interview study with 22 interviewees from teams of bio-medi...

The rapid advancement of artificial intelligence (AI) is changing our lives in many ways. One application domain is data science. New techniques in automating the creation of AI, known as AutoAI or AutoML, aim to automate the work practices of data scientists. AutoAI systems are capable of autonomously ingesting and pre-processing data, engineering...

By 2019, diversity is an established fact in most workplaces, teams, and work-groups, presenting both old and new challenges to CSCW in terms of team structure and technological supports for increasingly diverse teams. The research literature on diversity and teams has examined many definitions and attributes of diversity, and has described differe...

1 As Machine Learning (ML) systems become increasingly ubiquitous, capable and autonomous, it has become essential to take a human-centered view to consider how people's interactions with ML systems, including the effort to develop and evolve ML systems, impact their work practices, wellbeing and the social-organizational environment. Built on our...

HCI has a growing body of work regarding important social and community issues, as well as various grassroots communities working to make CHI more international and inclusive. In this workshop, we will build on this work: first reflecting on the contemporary CHI climate, and then developing an actionable plan towards making CHI2019 and subsequent S...

With the rise of big data, there has been an increasing need for practitioners in this space and an increasing opportunity for researchers to understand their workflows and design new tools to improve it. Data science is often described as data-driven, comprising unambiguous data and proceeding through regularized steps of analysis. However, this v...