To “In-House” or To Outsource? Artificial Intelligence in Canadian Local Governments
Why this work is in the frame
A frame that forgets how it found something cannot be audited. These are the routes that admitted this work.
Bibliographic record
Abstract
Artificial Intelligence (AI) promises significant benefits to municipalities, such as improved service delivery and operational efficiency. Given the resource-intensive nature of AI, municipalities must decide whether to develop AI in-house or outsource AI systems. Whereas outsourcing decisions have been widely studied in the private sector there is limited research to help us understand how municipalities navigate this choice despite their unique constraints, public accountability, and policy considerations. Our study addresses this gap by surveying representatives of 28 Canadian municipal AI projects to investigate the factors influencing their decisions. We found six factors that influence the decision of municipalities to “in-house” or outsource AI. We found that insufficient in-house AI expertise was the primary determinant of outsourcing. Funding and data sharing issues challenged both in-house development and outsourcing. Ensuring AI explainability and trust is perceived as more challenging when outsourcing. Contrary to common assumptions, AI maintenance is perceived as more difficult when outsourced. The lack of AI-specific regulations poses challenges for in-house development due to limited government guidance but also offers flexibility, while creating challenges in constructing AI outsourcing contracts. This paper is the first to compare in-house AI development and outsourcing within local governments. By capturing firsthand experiences of the participants directly involved in the AI projects, our research provides empirical insights into the trade-offs between these two approaches. Overall, these findings offer valuable guidance for municipalities seeking to make informed AI adoption decisions.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
Full frame distilled prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. Learned from the 10,348 direct Codex labels and 10,348 direct Gemma labels. Candidate is the union of thresholded teacher heads; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels or direct frontier model labels.
Codex and Gemma teacher scores by category
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.000 | 0.000 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.000 |
| Bibliometrics | 0.001 | 0.002 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.000 | 0.000 |
| Open science | 0.000 | 0.000 |
| Research integrity | 0.000 | 0.000 |
| Insufficient payload (model declined to judge) | 0.000 | 0.000 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from it