The growth of technologies promising to infer emotions raises political and ethical concerns, including concerns regarding their accuracy and transparency. A marginalized perspective in these conversations is that of data subjects potentially affected by emotion recognition. Taking social media as one emotion recognition deployment context, we conducted interviews with data subjects (i.e., social media users) to investigate their notions about accuracy and transparency in emotion recognition and interrogate stated attitudes towards these notions and related folk theories. We find that data subjects see accurate inferences as uncomfortable and as threatening their agency, pointing to privacy and ambiguity as desired design principles for social media platforms. While some participants argued that contemporary emotion recognition must be accurate, others raised concerns about possibilities for contesting the technology and called for better transparency. Furthermore, some challenged the technology altogether, highlighting that emotions are complex, relational, performative, and situated. In interpreting our findings, we identify new folk theories about accuracy and meaningful transparency in emotion recognition. Overall, our analysis shows an unsatisfactory status quo for data subjects that is shaped by power imbalances and a lack of reflexivity and democratic deliberation within platform governance.
more »
« less
Data Subjects' Conceptualizations of and Attitudes Toward Automatic Emotion Recognition-Enabled Wellbeing Interventions on Social Media
Automatic emotion recognition (ER)-enabled wellbeing interventions use ER algorithms to infer the emotions of a data subject (i.e., a person about whom data is collected or processed to enable ER) based on data generated from their online interactions, such as social media activity, and intervene accordingly. The potential commercial applications of this technology are widely acknowledged, particularly in the context of social media. Yet, little is known about data subjects' conceptualizations of and attitudes toward automatic ER-enabled wellbeing interventions. To address this gap, we interviewed 13 US adult social media data subjects regarding social media-based automatic ER-enabled wellbeing interventions. We found that participants' attitudes toward automatic ER-enabled wellbeing interventions were predominantly negative. Negative attitudes were largely shaped by how participants compared their conceptualizations of Artificial Intelligence (AI) to the humans that traditionally deliver wellbeing support. Comparisons between AI and human wellbeing interventions were based upon human attributes participants doubted AI could hold: 1) helpfulness and authentic care; 2) personal and professional expertise; 3) morality; and 4) benevolence through shared humanity. In some cases, participants' attitudes toward automatic ER-enabled wellbeing interventions shifted when participants conceptualized automatic ER-enabled wellbeing interventions' impact on others, rather than themselves. Though with reluctance, a minority of participants held more positive attitudes toward their conceptualizations of automatic ER-enabled wellbeing interventions, citing their potential to benefit others: 1) by supporting academic research; 2) by increasing access to wellbeing support; and 3) through egregious harm prevention. However, most participants anticipated harms associated with their conceptualizations of automatic ER-enabled wellbeing interventions for others, such as re-traumatization, the spread of inaccurate health information, inappropriate surveillance, and interventions informed by inaccurate predictions. Lastly, while participants had qualms about automatic ER-enabled wellbeing interventions, we identified three development and delivery qualities of automatic ER-enabled wellbeing interventions upon which their attitudes toward them depended: 1) accuracy; 2) contextual sensitivity; and 3) positive outcome. Our study is not motivated to make normative statements about whether or how automatic ER-enabled wellbeing interventions should exist, but to center voices of the data subjects affected by this technology. We argue for the inclusion of data subjects in the development of requirements for ethical and trustworthy ER applications. To that end, we discuss ethical, social, and policy implications of our findings, suggesting that automatic ER-enabled wellbeing interventions imagined by participants are incompatible with aims to promote trustworthy, socially aware, and responsible AI technologies in the current practical and regulatory landscape in the US.
more »
« less
- Award ID(s):
- 2020872
- NSF-PAR ID:
- 10437759
- Date Published:
- Journal Name:
- Proceedings of the ACM on Human-Computer Interaction
- Volume:
- 5
- Issue:
- CSCW2
- ISSN:
- 2573-0142
- Page Range / eLocation ID:
- 1 to 34
- Format(s):
- Medium: X
- Sponsoring Org:
- National Science Foundation
More Like this
-
-
Background Social networks such as Twitter offer the clinical research community a novel opportunity for engaging potential study participants based on user activity data. However, the availability of public social media data has led to new ethical challenges about respecting user privacy and the appropriateness of monitoring social media for clinical trial recruitment. Researchers have voiced the need for involving users’ perspectives in the development of ethical norms and regulations. Objective This study examined the attitudes and level of concern among Twitter users and nonusers about using Twitter for monitoring social media users and their conversations to recruit potential clinical trial participants. Methods We used two online methods for recruiting study participants: the open survey was (1) advertised on Twitter between May 23 and June 8, 2017, and (2) deployed on TurkPrime, a crowdsourcing data acquisition platform, between May 23 and June 8, 2017. Eligible participants were adults, 18 years of age or older, who lived in the United States. People with and without Twitter accounts were included in the study. Results While nearly half the respondents—on Twitter (94/603, 15.6%) and on TurkPrime (509/603, 84.4%)—indicated agreement that social media monitoring constitutes a form of eavesdropping that invades their privacy, over one-third disagreed and nearly 1 in 5 had no opinion. A chi-square test revealed a positive relationship between respondents’ general privacy concern and their average concern about Internet research (P<.005). We found associations between respondents’ Twitter literacy and their concerns about the ability for researchers to monitor their Twitter activity for clinical trial recruitment (P=.001) and whether they consider Twitter monitoring for clinical trial recruitment as eavesdropping (P<.001) and an invasion of privacy (P=.003). As Twitter literacy increased, so did people’s concerns about researchers monitoring Twitter activity. Our data support the previously suggested use of the nonexceptionalist methodology for assessing social media in research, insofar as social media-based recruitment does not need to be considered exceptional and, for most, it is considered preferable to traditional in-person interventions at physical clinics. The expressed attitudes were highly contextual, depending on factors such as the type of disease or health topic (eg, HIV/AIDS vs obesity vs smoking), the entity or person monitoring users on Twitter, and the monitored information. Conclusions The data and findings from this study contribute to the critical dialogue with the public about the use of social media in clinical research. The findings suggest that most users do not think that monitoring Twitter for clinical trial recruitment constitutes inappropriate surveillance or a violation of privacy. However, researchers should remain mindful that some participants might find social media monitoring problematic when connected with certain conditions or health topics. Further research should isolate factors that influence the level of concern among social media users across platforms and populations and inform the development of more clear and consistent guidelines.more » « less
-
Between 2018 and 2021 PIs for National Science Foundation Awards # 1758781 and 1758814 EAGER: Collaborative Research: Developing and Testing an Incubator for Digital Entrepreneurship in Remote Communities, in partnership with the Tanana Chiefs Conference, the traditional tribal consortium of the 42 villages of Interior Alaska, jointly developed and conducted large-scale digital and in-person surveys of multiple Alaskan interior communities. The survey was distributed via a combination of in-person paper surveys, digital surveys, social media links, verbal in-person interviews and telephone-based responses. Analysis of this measure using SAS demonstrated the statistically significant need for enhanced digital infrastructure and reworked digital entrepreneurial and technological education in the Tanana Chiefs Conference region. 1. Two statistical measures were created during this research: Entrepreneurial Readiness (ER) and Digital Technology needs and skills (DT), both of which showed high measures of internal consistency (.89, .81). 2. The measures revealed entrepreneurial readiness challenges and evidence of specific addressable barriers that are currently preventing (serving as hindrances) to regional digital economic activity. The survey data showed statistically significant correlation with the mixed-methodological in-person focus groups and interview research conducted by the PIs and TCC collaborators in Hughes and Huslia, AK, which further corroborated stated barriers to entrepreneurship development in the region. 3. Data generated by the survey and fieldwork is maintained by the Tanana Chiefs Conference under data sovereignty agreements. The survey and focus group data contains aggregated statistical/empirical data as well as qualitative/subjective detail that runs the risk of becoming personally identifiable especially due to (but not limited to) to concerns with exceedingly small Arctic community population sizes. 4. This metadata is being provided in order to serve as a record of the data collection and analysis conducted, and also to share some high-level findings that, while revealing no personal information, may be helpful for policymaking, regional planning and efforts towards educational curricular development and infrastructural investment. The sample demographics consist of 272 women, 79 men, and 4 with gender not indicated as a response. Barriers to Entrepreneurial Readiness were a component of the measure. Lack of education is the #1 barrier, followed closely by lack of access to childcare. Among women who participated in the survey measure, 30% with 2 or more children report lack of childcare to be a significant barrier to entrepreneurial and small business activity. For entrepreneurial readiness and digital economy, the scales perform well from a psychometric standpoint. The summary scores are roughly normally distributed. Cronbach’s alphas are greater than 0.80 for both. They are moderately correlated with each other (r = 0.48, p &amp;lt; .0001). Men and women do not differ significantly on either measure. Education is significantly related to the digital economy measure. The detail provided in the survey related to educational needs enabled optimized development of the Incubator for Digital Entrepreneurship in Remote Communities. Enhanced digital entrepreneurship training with clear cultural linkages to traditions and community needs, along with additional childcare opportunities are two among several specific recommendations provided to the TCC. The project PIs are working closely with the TCC administration and community members related to elements of culturally-aligned curricular development that respects data tribal sovereignty, local data management protocols, data anonymity and adherence to human subjects (IRB) protocols. While the survey data is currently embargoed and unable to be submitted publicly for reasons of anonymity, the project PIs are working with the NSF Arctic Data Center towards determining pathways for sharing personally-protected data with the larger scientific community. These approaches may consist of aggregating and digitally anonymizing sensitive data in ways that cannot be de-aggregated and that meet agency and scientific community needs (while also fully respecting and protecting participants’ rights and personal privacy). At present the data sensitivity protocols are not yet adapted to TCC requirements and the datasets will remain in their care.more » « less
-
null (Ed.)The DeepLearningEpilepsyDetectionChallenge: design, implementation, andtestofanewcrowd-sourced AIchallengeecosystem Isabell Kiral*, Subhrajit Roy*, Todd Mummert*, Alan Braz*, Jason Tsay, Jianbin Tang, Umar Asif, Thomas Schaffter, Eren Mehmet, The IBM Epilepsy Consortium◊ , Joseph Picone, Iyad Obeid, Bruno De Assis Marques, Stefan Maetschke, Rania Khalaf†, Michal Rosen-Zvi† , Gustavo Stolovitzky† , Mahtab Mirmomeni† , Stefan Harrer† * These authors contributed equally to this work † Corresponding authors: rkhalaf@us.ibm.com, rosen@il.ibm.com, gustavo@us.ibm.com, mahtabm@au1.ibm.com, sharrer@au.ibm.com ◊ Members of the IBM Epilepsy Consortium are listed in the Acknowledgements section J. Picone and I. Obeid are with Temple University, USA. T. Schaffter is with Sage Bionetworks, USA. E. Mehmet is with the University of Illinois at Urbana-Champaign, USA. All other authors are with IBM Research in USA, Israel and Australia. Introduction This decade has seen an ever-growing number of scientific fields benefitting from the advances in machine learning technology and tooling. More recently, this trend reached the medical domain, with applications reaching from cancer diagnosis [1] to the development of brain-machine-interfaces [2]. While Kaggle has pioneered the crowd-sourcing of machine learning challenges to incentivise data scientists from around the world to advance algorithm and model design, the increasing complexity of problem statements demands of participants to be expert data scientists, deeply knowledgeable in at least one other scientific domain, and competent software engineers with access to large compute resources. People who match this description are few and far between, unfortunately leading to a shrinking pool of possible participants and a loss of experts dedicating their time to solving important problems. Participation is even further restricted in the context of any challenge run on confidential use cases or with sensitive data. Recently, we designed and ran a deep learning challenge to crowd-source the development of an automated labelling system for brain recordings, aiming to advance epilepsy research. A focus of this challenge, run internally in IBM, was the development of a platform that lowers the barrier of entry and therefore mitigates the risk of excluding interested parties from participating. The challenge: enabling wide participation With the goal to run a challenge that mobilises the largest possible pool of participants from IBM (global), we designed a use case around previous work in epileptic seizure prediction [3]. In this “Deep Learning Epilepsy Detection Challenge”, participants were asked to develop an automatic labelling system to reduce the time a clinician would need to diagnose patients with epilepsy. Labelled training and blind validation data for the challenge were generously provided by Temple University Hospital (TUH) [4]. TUH also devised a novel scoring metric for the detection of seizures that was used as basis for algorithm evaluation [5]. In order to provide an experience with a low barrier of entry, we designed a generalisable challenge platform under the following principles: 1. No participant should need to have in-depth knowledge of the specific domain. (i.e. no participant should need to be a neuroscientist or epileptologist.) 2. No participant should need to be an expert data scientist. 3. No participant should need more than basic programming knowledge. (i.e. no participant should need to learn how to process fringe data formats and stream data efficiently.) 4. No participant should need to provide their own computing resources. In addition to the above, our platform should further • guide participants through the entire process from sign-up to model submission, • facilitate collaboration, and • provide instant feedback to the participants through data visualisation and intermediate online leaderboards. The platform The architecture of the platform that was designed and developed is shown in Figure 1. The entire system consists of a number of interacting components. (1) A web portal serves as the entry point to challenge participation, providing challenge information, such as timelines and challenge rules, and scientific background. The portal also facilitated the formation of teams and provided participants with an intermediate leaderboard of submitted results and a final leaderboard at the end of the challenge. (2) IBM Watson Studio [6] is the umbrella term for a number of services offered by IBM. Upon creation of a user account through the web portal, an IBM Watson Studio account was automatically created for each participant that allowed users access to IBM's Data Science Experience (DSX), the analytics engine Watson Machine Learning (WML), and IBM's Cloud Object Storage (COS) [7], all of which will be described in more detail in further sections. (3) The user interface and starter kit were hosted on IBM's Data Science Experience platform (DSX) and formed the main component for designing and testing models during the challenge. DSX allows for real-time collaboration on shared notebooks between team members. A starter kit in the form of a Python notebook, supporting the popular deep learning libraries TensorFLow [8] and PyTorch [9], was provided to all teams to guide them through the challenge process. Upon instantiation, the starter kit loaded necessary python libraries and custom functions for the invisible integration with COS and WML. In dedicated spots in the notebook, participants could write custom pre-processing code, machine learning models, and post-processing algorithms. The starter kit provided instant feedback about participants' custom routines through data visualisations. Using the notebook only, teams were able to run the code on WML, making use of a compute cluster of IBM's resources. The starter kit also enabled submission of the final code to a data storage to which only the challenge team had access. (4) Watson Machine Learning provided access to shared compute resources (GPUs). Code was bundled up automatically in the starter kit and deployed to and run on WML. WML in turn had access to shared storage from which it requested recorded data and to which it stored the participant's code and trained models. (5) IBM's Cloud Object Storage held the data for this challenge. Using the starter kit, participants could investigate their results as well as data samples in order to better design custom algorithms. (6) Utility Functions were loaded into the starter kit at instantiation. This set of functions included code to pre-process data into a more common format, to optimise streaming through the use of the NutsFlow and NutsML libraries [10], and to provide seamless access to the all IBM services used. Not captured in the diagram is the final code evaluation, which was conducted in an automated way as soon as code was submitted though the starter kit, minimising the burden on the challenge organising team. Figure 1: High-level architecture of the challenge platform Measuring success The competitive phase of the "Deep Learning Epilepsy Detection Challenge" ran for 6 months. Twenty-five teams, with a total number of 87 scientists and software engineers from 14 global locations participated. All participants made use of the starter kit we provided and ran algorithms on IBM's infrastructure WML. Seven teams persisted until the end of the challenge and submitted final solutions. The best performing solutions reached seizure detection performances which allow to reduce hundred-fold the time eliptologists need to annotate continuous EEG recordings. Thus, we expect the developed algorithms to aid in the diagnosis of epilepsy by significantly shortening manual labelling time. Detailed results are currently in preparation for publication. Equally important to solving the scientific challenge, however, was to understand whether we managed to encourage participation from non-expert data scientists. Figure 2: Primary occupation as reported by challenge participants Out of the 40 participants for whom we have occupational information, 23 reported Data Science or AI as their main job description, 11 reported being a Software Engineer, and 2 people had expertise in Neuroscience. Figure 2 shows that participants had a variety of specialisations, including some that are in no way related to data science, software engineering, or neuroscience. No participant had deep knowledge and experience in data science, software engineering and neuroscience. Conclusion Given the growing complexity of data science problems and increasing dataset sizes, in order to solve these problems, it is imperative to enable collaboration between people with differences in expertise with a focus on inclusiveness and having a low barrier of entry. We designed, implemented, and tested a challenge platform to address exactly this. Using our platform, we ran a deep-learning challenge for epileptic seizure detection. 87 IBM employees from several business units including but not limited to IBM Research with a variety of skills, including sales and design, participated in this highly technical challenge.more » « less
-
To enable targeted ads, companies profile Internet users, automatically inferring potential interests and demographics. While current profiling centers on users' web browsing data, smartphones and other devices with rich sensing capabilities portend profiling techniques that draw on methods from ubiquitous computing. Unfortunately, even existing profiling and ad-targeting practices remain opaque to users, engendering distrust, resignation, and privacy concerns. We hypothesized that making profiling visible at the time and place it occurs might help users better understand and engage with automatically constructed profiles. To this end, we built a technology probe that surfaces the incremental construction of user profiles from both web browsing and activities in the physical world. The probe explores transparency and control of profile construction in real time. We conducted a two-week field deployment of this probe with 25 participants. We found that increasing the visibility of profiling helped participants anticipate how certain actions can trigger specific ads. Participants' desired engagement with their profile differed in part based on their overall attitudes toward ads. Furthermore, participants expected algorithms would automatically determine when an inference was inaccurate, no longer relevant, or off-limits. Current techniques typically do not do this. Overall, our findings suggest that leveraging opportunistic moments within pervasive computing to engage users with their own inferred profiles can create more trustworthy and positive experiences with targeted ads.more » « less
- Free Publicly Accessible Full Text
-
Accepted Manuscript1.0
- Journal Article:
- https://doi.org/10.1145/3476049
