.

photo of pink flowers

one aside based on @rawsignal.ca:

I once heard it touted that AI would level the field for marginalised people, by ensuring people who'd never created a pitch deck/cover letter could make one now.

if VCs are no longer reading pitch decks because they're all AI, then they're going on "gut" = bias

How access to information can help us understand AI decision making Dive into the world of AI decision-making and discover how information access plays a key role in understanding this new kind of governance.

Join us on Tuesday 19th for our next FOI in practice webinar, all about government use of AI in decision making

www.eventbrite.co.uk/e/how-access...

None of this is real and it doesn't matter - Raw Signal Group The difference between the theater of work and actual work.

The difference between the theater of work and actual work.

www.rawsignal.ca/newsletter-a...

Engineer restores pay phones for free public use In a rural Vermont county, spotty cellphone coverage is the norm. But a local tinkerer is using old technology to help his community stay connected free of charge.

Love this story www.npr.org/2025/08/04/n...

How Social Media Shortens Your Life And How to Expand it

“Chatbots are also becoming curvilinear, increasingly ending their answers with a question or an offer for further help to create a kind of conversational infinite scroll.”

www.gurwinder.blog/p/how-social...

Public policymaking: from AI to decomputing AI seems to offer many benefits to public policymaking, but it can't address the tricky structural issues that impede actual change.

"A pivot to decomputing is a way to reassert the value of situated knowledge and of context over scale." @danmcquillan.bsky.social

www.jrf.org.uk/ai-for-publi...

also are there any suggestions for self-hosted/created LMS type set ups so I can opt out of eclass? I really don’t have the time or infrastructural knowledge to build a bespoke thing but I need to do something to get out of that system

I don't usually read dissertations - this will be one I do read (and I've already read excellent parts of it!)

Come hang out with me at Civic Spark next weekend!!

full heart after @civictechto.bsky.social - you should join us sometime, all are welcome <3

Design versus reality: assessing the results and compliance of algorithmic impact assessments - Digital Society Algorithmic impact assessments (AIAs) have become a dominant regulatory instrument in governing artificial intelligence (AI). While there are noteworthy examples across the global north, Canada’s AIA is considered to be the best practice worldwide. When AIAs are studied, evaluations have been based on the assessment of the instrument and not the examination of their answers. We examine Canada’s published AIAs. We report five findings: (1) Uneven compliance is observed in the completion of AIAs; (2) Reasons for automation legitimize efficiency and innovation narratives; (3) Impacts and trade-offs are framed as non-existent, positive and undermine harms; (4) Civil society organizations are non-existent in AIAs; and (5) Accountability is framed as processual mitigation of AI impacts. Despite the promise of AIAs for accountability of AI systems, our results reveal a “design-reality” gap between literature and practice. We observed that any negative impacts were framed positively; input was not elicited from the public; and an over-emphasis of self-regulation conformed to organizational procedures instead of investigating outcomes. Although submission is mandatory, its processual accountability failed to ensure compliance. We recommend strengthening accountability to include civil society, formalizing harms instead of emphasizing impacts or risks and blending processual accountability with outcomes of AI systems.

Unfortunately the paper is not open access. Send me a message or email if you cannot access it and would like a copy!

link.springer.com/article/10.1...

An Analysis of Canada's Algorithmic Impact Assessments Hosted on the Open Science Framework

We also created an open repository of the AIAs, which we hope will encourage additional research on them. The repository is available here: osf.io/rk8ux/

1 Uneven compliance is observed in completion of AIAs
2 Reasons for automation legitimize efficiency & innovation narratives
3 Impacts & trade-offs are framed as non-existent, positive & undermine harms
4 CSOs are non-existent in AIAs
5 Accountability framed as processual mitigation of AI impacts

After more than two years of presentations, discussions and analysis, @resieber.bsky.social and I are excited to share the first empirical paper on published algorithmic impact assessments (AIAs).

We examine 26 AIAs based on AI systems used by the Canadian government and report five findings:

Design versus reality: assessing the results and compliance of algorithmic impact assessments - Digital Society Algorithmic impact assessments (AIAs) have become a dominant regulatory instrument in governing artificial intelligence (AI). While there are noteworthy examples across the global north, Canada’s AIA is considered to be the best practice worldwide. When AIAs are studied, evaluations have been based on the assessment of the instrument and not the examination of their answers. We examine Canada’s published AIAs. We report five findings: (1) Uneven compliance is observed in the completion of AIAs; (2) Reasons for automation legitimize efficiency and innovation narratives; (3) Impacts and trade-offs are framed as non-existent, positive and undermine harms; (4) Civil society organizations are non-existent in AIAs; and (5) Accountability is framed as processual mitigation of AI impacts. Despite the promise of AIAs for accountability of AI systems, our results reveal a “design-reality” gap between literature and practice. We observed that any negative impacts were framed positively; input was not elicited from the public; and an over-emphasis of self-regulation conformed to organizational procedures instead of investigating outcomes. Although submission is mandatory, its processual accountability failed to ensure compliance. We recommend strengthening accountability to include civil society, formalizing harms instead of emphasizing impacts or risks and blending processual accountability with outcomes of AI systems.

📢 New Research Paper! Design versus reality: assessing the results and compliance of algorithmic impact assessments

link.springer.com/article/10.1...

interesting. thank you

DRAFT — Tech and Toronto City Budget Civic Tech Project Origin Story and Direction — (Also known as, welcome, please join us, start here)

some background:

biancawylie.medium.com/draft-tech-a...

:)

join us tonight? :)

when we do breakouts, will be doing one about slowly reviving/revising a project to try to figure out ways to map public tech infrastructure.

nice / new to me too, if you do will like to know what you think :)

Fall Novel Workshops street art by Shanalee Hampton Hey folks, I hope that your summer is going well! I’ve been working with some great writers in coaching sessions and manuscript consultations, and will now take Augus…

Are you looking for a novel workshop? I've got you.
cariluna.com/fall-novel-w...

Public policymaking: from AI to decomputing AI seems to offer many benefits to public policymaking, but it can't address the tricky structural issues that impede actual change.

decomputing is the innovation agenda

www.jrf.org.uk/ai-for-publi...

Abstract page from a dissertation by me entitled "The Impacts and Governance of Artificial Intelligence: A Transfeminist Perspective." The excerpted abstract reads: "This thesis investigates the impacts and governance of artificial intelligence (AI) systems through a transfeminist lens, focusing analysis upon challenges of power, exclusion, and injustice alongside opportunities for advancing equity, community-based resistance, and transformative change. AI governance is a field of research and practice seeking to maximize benefits and minimize harms caused by AI systems. However, AI governance is frequently ineffective at preventing AI systems from causing harm to society and the environment, with historically marginalized groups being particularly vulnerable to harm. Applying a framework of theories drawn from service science, feminist studies, and trans studies, I analyze relationships between AI governance and harm prevention in three separate co-authored, peer-reviewed articles. The first article develops a theory linking beneficial and harmful impacts caused by AI systems to the value chains through which various actors integrate resources and co-create value throughout the AI system lifecycle. This theory is applied to an integrative review of ethical concerns implicated in AI systems and to discuss future directions for intervening in the impacts caused by AI systems. The second article presents a semi-systematic review and content analysis of 84 AI governance initiatives launched by federal and provincial governments in Canada from 2017 to 2022. AI governance initiatives are used to organize many types of interventions, and Canada’s initiatives favor intervention in the impacts of AI on Canadian industry, innovation,"

My dissertation is now available online! The diss applies trans and feminist theory to analyze structural injustices in 3 of my empirical studies of AI governance & policymaking. Give it a look if you're concerned about power dynamics in AI policy

www.proquest.com/docview/3234...

working on a windows 95 themed automation game about making powerpoint decks #gamedev

Decidim and the Spirit of a Digital Commons Community This second interview is with Decidim, which is leading the charge in showing how Digital Commons can work at scale as infrastructure. Decidim is an open source participatory democracy platform that e...

“Without the public sector, we cannot understand Decidim. It would never exist without the Barcelona City Council. They had to really believe in us, in how to build Decidim as not just a piece of free software, but a Digital Common. This was crucial.”

commons.ngi.eu/2025/08/04/d...

ye olde is this or isn't this a new shape of a problem. annnd with that, good night :) (it's usually not)

there's something here about how a highly regulated profession is already used to having to pay v contextual and specific attention to take care with individual patient data. seems less clear as to how the profession is or isn't used to having to think about tech systems at the macro level.

AI-powered tool on St. Michael’s surgical unit to improve patient care On a busy day in St. Michael’s Hospital’s surgical unit, the care team can look after upwards of 42 patients, coordinating everything from pain management to wound treatment to deciding when patients…

there have been a few instances of different pooled health data projects that i want to spend more time on - this chartwatch thing flared up in the cbc a while ago and i've lost track of the data it was drawing from unityhealth.to/2023/02/char...

the austerity and automation rationale loop is awful