Lightnews — Scholar-powered news

Reposted by Julian Skirzynski

Dr. Damien P. Williams Wants A Better Sociotechnical Paradigm @wolvendamien.bsky.social · 3d

Preliminary results show that the current framework of "AI" makes ppl less likely to help or seek help from other humans, or to seek to soothe conflict, and that people actively prefer that framework to any others, literally serving to make them more dependent on it.

Sycophantic AI Decreases Prosocial Intentions and Promotes Dependence

Both the general public and academic communities have raised concerns about sycophancy, the phenomenon of artificial intelligence (AI) excessively agreeing with or flattering users. Yet, beyond isolat...

arxiv.org

11 120 250

Reposted by Julian Skirzynski

Ezequiel Lopez-Lopez @eloplop.bsky.social · Jul 28

New research out!🚨

In our new paper, we discuss how generative AI (GenAI) tools like ChatGPT can mediate confirmation bias in health information seeking.
As people turn to these tools for health-related queries, new risks emerge.
🧵👇
nyaspubs.onlinelibrary.wiley.com/doi/10.1111/...

NYAS Publications

Generative artificial intelligence (GenAI) applications, such as ChatGPT, are transforming how individuals access health information, offering conversational and highly personalized interactions. Whi...

nyaspubs.onlinelibrary.wiley.com

1 9 15

Julian Skirzynski @jskirzynski.bsky.social · Jun 24

Sure :)

Julian Skirzynski @jskirzynski.bsky.social · Jun 24

We’ll be presenting ‪@facct on 06.24 at 10:45 AM during the Evaluating Explainable AI session!

Come chat with us. We would love to discuss implications for AI policy, better auditing methods, and next steps for algorithmic fairness research.

#AIFairness #XAI

1

Julian Skirzynski @jskirzynski.bsky.social · Jun 24

But if they are indeed used to dispute discrimination claims, we can expect multiple failed cases due to insufficient evidence and many undetected discriminatory decisions.

Current explanation-based auditing is, therefore, fundamentally flawed, and we need additional safeguards.

1 1

Julian Skirzynski @jskirzynski.bsky.social · Jun 24

Despite their unreliability, explanations are suggested as anti-discrimination measures by a number of regulations.

GDPR ✓ Digital Services Act ✓ Algorithmic Accountability Act ✓ GDPD (Brazil) ✓

1 1

Julian Skirzynski @jskirzynski.bsky.social · Jun 24

So why do explanations fail?

1️⃣ They target individuals, while discrimination operates on groups
2️⃣ Users’ causal models are flawed
3️⃣ Users overestimate proxy strength and treat its presence in the explanation as discrimination
4️⃣ Feature-outcome relationships bias user claims

1 1

Julian Skirzynski @jskirzynski.bsky.social · Jun 24

BADLY.

When participants flag discrimination, they are correct ~50% of the time, miss 55% of the discriminatory predictions and keep a 30% FPR.

Additional knowledge (protected attributes, proxy strength) improves the detection to roughly 60% without affecting other measures.

1 2

Julian Skirzynski @jskirzynski.bsky.social · Jun 24

Our setup lets us assign each robot a ground-truth discrimination outcome, which lets us evaluate how well each participant could do under different information regimes.

So, how did they do?

1 1

Julian Skirzynski @jskirzynski.bsky.social · Jun 24

We recruited participants, anchored their beliefs on discrimination, trained them to use explanations, and tested to make sure they got it right.

We then saw how well they could flag unfair predictions based on counterfactual explanations and feature attribution scores.

1 1

Julian Skirzynski @jskirzynski.bsky.social · Jun 24

Participants audit a model to predict if robots sent to Mars will break down. Some are built by “Company X.” Others by “Company S.”

Our model predicts failure based on robot body parts. It can discriminate against Company X by predicting that robots without an antenna fail.

1 1

Julian Skirzynski @jskirzynski.bsky.social · Jun 24

We cannot tell if explanations work or not due to these reasons.

To tackle this challenge, we introduce a synthetic task where we:
- Teach users how to use explanations
- Control their beliefs
- Adapt the world to fit their beliefs
- Control the explanation content

1 1

Julian Skirzynski @jskirzynski.bsky.social · Jun 24

Users may fail to detect discrimination through explanations due to:

- Proxies not being revealed by explanations
- Issues with interpreting explanations
- Wrong assumptions about proxy strength
- Unknown protected class
- Incorrect causal beliefs

1 2

Julian Skirzynski @jskirzynski.bsky.social · Jun 24

Imagine a model that predicts loan approval based on credit history and salary.

Would a rejected female applicant get approved if she somehow applied as a man?

If yes, her prediction was discriminatory.

Fairness requires predictions to stay the same regardless of the protected class.

1 1

Julian Skirzynski @jskirzynski.bsky.social · Jun 24

Right to explanation laws assume explanations help people detect algorithmic discrimination.

But is there any evidence for that?

In our latest work w/ David Danks @berkustun, we show explanations fail to help people, even under optimal conditions.

PDF shorturl.at/yaRua

1 8

Julian Skirzynski @jskirzynski.bsky.social · Jun 11

You're both in!

1

Reposted by Julian Skirzynski

Harry Cheon @scheon.com · Apr 24

Denied a loan, an interview, or an insurance claim by machine learning models? You may be entitled to a list of reasons.

In our latest w @anniewernerfelt.bsky.social @berkustun.bsky.social @friedler.net, we show how existing explanation frameworks fail and present an alternative for recourse