However, it is important to note that this development raises important questions regarding governance, digital autonomy, privacy and security. Questions that the people we interviewed are aware of and aim to mitigate.
The report describes our insights - feel free to get in touch!
Our research into the use of generative AI by Dutch government organisations is out. The results show that the use of GenAI has increased by a lot. Reasons to use GenAI is the wish to increase efficiency and avoid shadow use of tools. Read here (in Dutch): www.tno.nl/nl/newsroom/...
This work was done in the context of the Diversifair Project, which is funded by Erasmus+ diversifair-project.eu
We identify three key characteristics of AI harms.
- Intentionality
- Temporality
- Interactivity
By untangling these characteristics, we can begin to address the structurality, systemic, or procedural elements that enable AI harms to emerge.
It's out! Our newest publication is published in Oxford Intersections. We focus on understanding AI harms through their real-world impacts and beyond a risk-based perspective.
Read it here: academic.oup.com/edited-volum...
๐ง Transformation: how trustworthiness is actively shaped during development.
๐จ๏ธ Revelation: how developers communicate to generate trust.
We also reflect on the challenges of developing (for) trust and trustworthiness and note that both practices need each other.
โจ New publication out!
We collaborated with AI developers to understand how they build trustworthiness into AI systems and how trust is generated. We found to overarching practices: Transformation and Revelation >
2. Trustworthy workplaces, trustworthy AI? A necessary diversity climate change for the AI domain - developing trustworthy AI is a complex task and teams benefits from incorporating multiple diverse perspectives. But is the AI domain ready for this?
Curious? Come to the sessions or get in touch!
Tonight #ICA25 kicks off and I'm happy to be a part of it!
I'll present to of my studies tomorrow:
1. โMore of an Academic Exerciseโ: A Critical Assessment of Sociotechnical AI Evaluation Practices (and Why They Fail) - about AI evaluations in R&D contexts (spoiler: more needs to be done!)
WAI Connect Zuid Holland - When AI Meets UX and Artifcial (im)perfection
On 20 May, I will discuss AI (im)perfection and share insights from my PhD research and experience as a researcher at TNO Vector. Everybody's welcome so feel free to sign up! Details here:
๐ www.linkedin.com/events/waico...
A stack of books and one upright copy, showing the cover of the dissertation titled "Practices we Value. A multimethodological exploration of trustworthy AI development."
Very proud of how my dissertation turned out, both in terms of content and looks! I look foward to my defence on 20 March and to being PhDone ๐
We're conducting a survey on interpretability/explainability. If you are a researcher in XAI or use XAI tools in practice, please help us by filling this short survey!
RTs appreciated!
docs.google.com/forms/d/e/1F...
๐Tackle data issues and black-box AI at the SPATIAL Project Final Event! Donโt miss insightful discussions, networking, and poster sessions.
๐
July 1st, 10 AM - 5 PM CET.
๐ Faculty of Technology, Policy & Management, TU Delft, Jaffalaan 5, 2628BX, Delft, NL
Secure your spot: lnkd.in/e4JA-apX
