The Montreal AI Ethics Institute Montreal AI Ethics Institute has published a concise summary of our paper on *bridging systems* — algorithmic systems (incl. recommenders) that help build mutual understanding and trust across divides.

montrealethics.ai/bridging-syste…

(Thanks Abhishek Gupta!)

Montreal AI Ethics Institute Montreal AI Ethics Institute just published a research summary of our #EAAMO23 paper on what participation means in the context of AI design and development!

Blog Post: twtr.to/4nO2E

Full Paper: dl.acm.org/doi/10.1145/36…

Fernando Alonso Delgado Michael Madaio Qian Yang

Thanks to Montreal AI Ethics Institute Montreal AI Ethics Institute for sharing the summary of our recent ICSA’24 paper on A Reference Architecture for Designing Foundation Model based Agents
montrealethics.ai/responsible-ge…
Paper preprint on arxiv: arxiv.org/abs/2311.13148
Abhishek Gupta

Thanks to Abhishek Gupta from Montreal AI Ethics Institute for featuring my research work about Generative AI in Writing Research Papers in Montreal AI Ethics Institute's research summaries institute.

Check it out here: montrealethics.ai/generative-ai-…

LLMs like GPT-4 are shaping our world. Can we agree on the principles that guide their reasoning?

We assembled a list of 37 such principles, based on ideas spanning from AI to philosophy.

Summary Montreal AI Ethics Institute
montrealethics.ai/a-collection-o…

Paper SoLaR @ NeurIPS2023
arxiv.org/abs/2312.10059

Brown CS research that recently won a SoLaR Best Paper Award (go.brown.edu/SoLaRBestPaper) has been making waves this week. Check out what they're saying at Montreal AI Ethics Institute (montrealethics.ai/low-resource-l…) and The Register (theregister.com/2024/01/31/gpt…).

Managing #AI #risks after model implementation ...
Some answers from @_IAPS_ Montreal AI Ethics Institute AI Newsletter

Missed @_IAPS_ report on Deployment Corrections? This 2-page summary written by Joe O’Brien and published by Montreal AI Ethics Institute explores how AI developers can prepare to decisively manage risks discovered after model deployment. montrealethics.ai/deployment-cor…

We need to rethink the belief that all privacy-preserving models are inherently more discriminatory.

I give a high-level overview of why - in this Montreal AI Ethics Institute blog post based on my summer internship work at Google DeepMind montrealethics.ai/unlocking-accu… 1/5

My research summary on defense against authorship identification attacks is now available on Montreal AI Ethics Institute🤗 Useful tips are included for whistleblowers💪🏻
montrealethics.ai/defending-agai…

Great to see World Privacy Forum's Risky Analysis report featured on Montreal AI Ethics Institute! It shows how governments around world are implementing responsible AI & need for evaluation of methods used to measure AI fairnss/explainability amid emerging FAULTY AI FIXES:
montrealethics.ai/risky-analysis…

The Montreal AI Ethics Institute Montreal AI Ethics Institute recently published a summary of one of my papers on topological aspects of sentence embeddings in GPT models (montrealethics.ai/do-large-gpt-m…). I visualize moral dimensions in these high dimensional representation manifolds by summarizing

Our ongoing research on the toxicity-rabbit hole framework reveals that it is not just PaLM 2, several other #LLM can produce extremely harmful text targeting minorities. Below are (censored) horrific #Antisemitic and #racist generative AI examples produced by models with varying

How safe are LLM guardrails in protecting disadvantaged groups? Turns out, not very much.

montrealethics.ai/down-the-toxic…

#aiforgood #safetyforall #aisafety #responsibleai

Excellent piece in Montreal AI Ethics Institute by Guru on our work on zero knowledge proofs of training (zkPOT). I really like that he connected zkPOT to the legal landscape. A paper on this appeared at CCS last month. Go Guru!

montrealethics.ai/experimenting-…

Research summary of our recent study on the political biases of ChatGPT is out in Montreal AI Ethics Institute! montrealethics.ai/ai-neutrality-…

ChatGPTの政治バイアスに関する最近の研究を
Montreal AI Ethics Instituteで紹介させて頂きました。AIは中立になっていっているのか、それとも見せかけか。montrealethics.ai/ai-neutrality-…

Thank you to Montreal AI Ethics Institute for covering our work on Chain of Density!

Alex Fabbri, Eric Lehman, Faisal Ladhak, Noémie Elhadad

montrealethics.ai/from-sparse-to…

A Machine Learning Challenge or a Computer Security Problem? | Montreal AI Ethics Institute montrealethics.ai/a-machine-lear…

david glukhov Nicolas Papernot Yarin

New research from UMSI:
“An Empirical Analysis of Racial Categories in the Algorithmic Fairness Literature”
Amina A. Abdu, Irene V. Pasquetto, Abigail Z. Jacobs
Amina Abdu Irene Pasquetto Abigail Jacobs
Montreal AI Ethics Institute