Automated Annotations for AI Data and Model Transparency

The data and Artificial Intelligence revolution has had a massive impact on enterprises, governments, and society alike. It is fueled by two key factors. First, data have become increasingly abundant and are often available openly. Enterprises have more data than they can process. Governments are spearheading open data initiatives by setting up data portals such as data.gov and releasing large amounts of data to the public. Second, AI engineering development is becoming increasingly democratized. Open source frameworks have enabled even an individual developer to engineer sophisticated AI systems. But with such ease of use comes the potential for irresponsible use of data. Ensuring that AI systems adhere to a set of ethical principles is one of the major problems of our age. We believe that data and model transparency has a key role to play in mitigating the deleterious effects of AI systems. In this article, we describe a framework to synthesize ideas from various domains such as data transparency, data quality, data governance among others to tackle this problem. Specifically, we advocate an approach based on automated annotations (of both data and the AI model), which has a number of appealing properties. The annotations could be used by enterprises to get visibility of potential issues, prepare data transparency reports, create and ensure policy compliance, and evaluate the readiness of data for diverse downstream AI applications. We propose a model architecture and enumerate its key components that could achieve these requirements. Finally, we describe a number of interesting challenges and opportunities.

The effect of perceived organizational culture on employees’ information security compliance

Purpose This paper aims to investigate the connection between different perceived organizational cultures and information security policy compliance among white-collar workers. Design/methodology/approach The survey using the Organizational Culture Assessment Instrument was sent to white-collar workers in Sweden (n = 674), asking about compliance with information security policies. The survey instrument is an operationalization of the Competing Values Framework that distinguishes between four different types of organizational culture: clan, adhocracy, market and bureaucracy. Findings The results indicate that organizational cultures with an internal focus are positively related to employees’ information security policy compliance. Differences in organizational culture with regards to control and flexibility seem to have less effect. The analysis shows that a bureaucratic form of organizational culture is most fruitful for fostering employees’ information security policy compliance. Research limitations/implications The results suggest that differences in organizational culture are important for employees’ information security policy compliance. This justifies further investigating the mechanisms linking organizational culture to information security compliance. Practical implications Practitioners should be aware that the different organizational cultures do matter for employees’ information security compliance. In businesses and the public sector, the authors see a development toward customer orientation and marketization, i.e. the opposite an internal focus, that may have negative ramifications for the information security of organizations. Originality/value Few information security policy compliance studies exist on the consequences of different organizational/information cultures.

I Am Better Than Others: Waste Management Policies and Self-Enhancement Bias

Waste source separation has been a social dilemma globally with a low participation rate. This research attempted to solve this dilemma by exploring the effect of mandatory (versus voluntary) policies on waste separation from the perspective of the self-versus based on deterrence theory and self-enhancement motivation. Hypothetical scenarios were used to demonstrate the effectiveness of mandatory policies and self-enhancement bias for residents (n = 589) and adolescents (n = 121). Study 2 was performed to replicate the findings of Study 1 with a no-implementation policy condition, and Study 3 extended the findings to adolescents. We found robust self-enhancement bias, where participants perceived themselves to be better than others in both willingness to perform and attitudes toward waste separation behavior. Specifically, participants tended to perceive themselves to perform waste separation well when policy compliance was voluntary, but they tended to perceive others to perform well when policy compliance was mandatory with supervision. These findings highlight the impact of mandatory policy with supervision and self-enhancement bias in waste management. The present studies provide substantial evidence and implications for the necessity of supervision in mandatory policy implementation.

Implementation of mandatory opioid prescribing limits in North Carolina: healthcare administrator and prescriber perspectives

Background Recent increases in state laws to reduce opioid prescribing have demonstrated a need to understand how they are interpreted and implemented in healthcare systems. The purpose of this study was to explore the systems, strategies, and resources that hospital administrators and prescribers used to implement the 2017 North Carolina Strengthen Opioid Prevention (STOP) Act opioid prescribing limits, which limited initial prescriptions to a five (for acute) or seven (for post-surgical) days’ supply. Methods We interviewed 14 hospital administrators and 38 prescribers with degrees in medicine, nursing, pharmacy, business administration and public health working across North Carolina. Interview guides, informed by the Consolidated Framework for Implementation Research, explored barriers and facilitators to implementation. Interview topics included communication, resources, and hospital system support. Interviews were recorded and transcribed, then analyzed using flexible coding, integrating inductive and deductive coding, to inform analytic code development and identify themes. Results We identified three main themes around implementation of STOP act mandated prescribing limits: organizational communication, prescriber education, and changes in the electronic medical record (EMR) systems. Administrators reflected on implementation in the context of raising awareness and providing reminders to facilitate changes in prescriber behavior, operationalized through email and in-person communications as well as dedicated resources to EMR changes. Prescribers noted administrative communications about prescribing limits often focused on legality, suggesting a directive of the organization’s policy rather than a passive reminder. Prescribers expressed a desire for more spaces to have their questions answered and resources for patient communications. While hospital administrators viewed compliance with the law as a priority, prescribers reflected on concerns for adequately managing their patients’ pain and limited time for clinical care. Conclusions Hospital administrators and prescribers approached implementation of the STOP act prescribing limits with different mindsets. While administrators were focused on policy compliance, prescribers were focused on their patients’ needs. Strategies to implement the mandate then had to balance patient needs with policy compliance. As states continue to legislate to prevent opioid overdose deaths, understanding how laws are implemented by healthcare systems and prescribers will improve their effectiveness through tailoring and maximizing available resources.