Algorithms and Power in Public Higher Education in Canada

Conceitos Básicos

The adoption of data-driven practices and algorithmic decision-making in public higher education institutions leads to increased student surveillance, exacerbation of existing inequities, and the automation of the faculty-student relationship.

Resumo

The article discusses the adoption of algorithmic decision-making in public higher education institutions.
It highlights the impact on students, including increased surveillance, exacerbation of existing inequities, and automation of relationships.
The study conducted a qualitative analysis at a public college in Ontario, Canada.
Data collection, algorithms used, outcomes, and stakeholder perceptions were assessed.
The study identified a cycle of increased institutional power perpetuated by algorithmic decision-making.

Estatísticas

"We found that these factors impact students through increased tracking and monitoring of learning activities."
"While algorithms are currently limited in use, the college plans to expand its machine learning capacity."
"The savings provided from cost-cutting are then reinvested into revenue-generating programs."

Citações

"Data is viewed as a way to cope with resource scarcity."
"The financial promises of automation through algorithmic decision-making go beyond just the allocation of limited resources."
"The emphasis on financial sustainability has made algorithmic and data-driven decision-making a top priority for the college."

Principais Insights Extraídos De

"This is not a data problem"

by Kelly McConv... às arxiv.org 03-22-2024

https://arxiv.org/pdf/2403.13969.pdf

Perguntas Mais Profundas

Why are student-data-driven and algorithmic decision-making used within higher education?

Student data-driven and algorithmic decision-making are utilized in higher education for several reasons. Firstly, there is a growing emphasis on accountability, efficiency, and effectiveness in post-secondary institutions driven by neoliberal governance strategies. These approaches prioritize metrics, key performance indicators (KPIs), and financial sustainability [43]. By leveraging student data and algorithms, institutions can make evidence-based decisions to support student success, retention, quality assurance, process improvements, revenue forecasting through enrollment prediction models [59], resource allocation management amidst resource scarcity [20], as well as potential commercialization opportunities of trained algorithms [29].

How can biases be mitigated in automated algorithmic systems?

Mitigating biases in automated algorithmic systems is crucial to ensure fair outcomes and prevent discrimination. Several strategies can be employed:

Diverse Data Representation: Ensuring that training datasets are diverse and representative of the population to avoid biased predictions.
Regular Auditing: Conducting regular audits of the algorithms to identify any bias or discriminatory patterns.
Transparency & Explainability: Making the decision-making process transparent so stakeholders understand how decisions are reached.
Bias Detection Algorithms: Implementing bias detection algorithms that flag potential biases during model development.
Ethical Guidelines & Oversight: Establishing clear ethical guidelines for algorithm design and deployment with oversight mechanisms.

How does the increasing reliance on data affect student autonomy?

The increasing reliance on data in educational institutions can have implications for student autonomy:

Increased Surveillance: Students may feel constantly monitored through tracking their activities leading to a loss of privacy.
Exacerbation of Inequities: Data-driven decisions may inadvertently reinforce existing inequalities based on historical data trends or systemic biases.
Automation of Student-Faculty Relationship: The use of algorithms could automate aspects of the faculty-student relationship potentially reducing personalized interactions between students and educators.

These factors combined could impact students' ability to make independent choices or have control over their educational experiences due to increased institutional monitoring and automation processes influenced by data-driven practices within higher education settings