Observer | The Legal and Ethical Minefield of A.I.-Driven Employee Surveillance by Kayvon Touran and Ben Dattner

AI generated summary, Read the full article for complete information.

The article warns that AI systems originally marketed as objective performance tools are now being used to continuously monitor, profile, and even manipulate employees, creating a new “surveillance wages” dynamic where pay and career decisions are driven by hidden algorithms that analyze keystrokes, email sentiment, location data, and even personal financial or health information. This deep‑level monitoring can amplify existing biases, erode privacy, and undermine legal protections such as the ADA, Title VII, and the NLRA because current employment laws were designed for human decision‑makers, not opaque algorithmic systems. The authors detail how companies—from major tech platforms like Microsoft’s Viva Insights to warehouse giants like Amazon—are deploying tools that infer productivity, psychological states, and financial vulnerability, then use those insights to shape compensation, promotions, and even employee behavior without consent. To curb these risks, they advocate transparency, employee consent, independent validation of AI models, human oversight of critical decisions, and proactive engagement with emerging regulations that aim to limit algorithmic wage‑setting and protect workers from covert manipulation.

Read more: https://observer.com/2026/05/legal-ethical-risks-ai-employee-profiling-workplace-monitoring/

#artificialintelligence #surveillancewages #employeesurveillance #algorithmicbias #privacyrights

Tubefilter: Social media has political divides, but some feeds are more polarized than others. “Researchers set up 323 ‘sock puppet’ accounts on TikTok to measure the political polarization of the For You Page, and they found some apparent disparities between right-leaning and left-leaning feeds.”

⚖️ When the Math Decides: Algorithms, Liberty & the Fight to Stay Human

Your childcare was cancelled by math. No human saw your file. An algorithm flagged your neighbourhood, a late bill & your roommate's parking ticket — and upended your life.

This is already happening. New episode of Heliox. 🧵

#AIandHumanRights #DigitalRights #AlgorithmicBias

RE: https://fediscience.org/@oatp/116438875358622213

"Maximal transparency is almost certainly not ethically desirable."

Desirable 'for whom'?

For platforms facing regulatory scrutiny, opacity is a feature. For users discriminated against by biased recommendation engines, transparency is survival. For communities targeted by algorithmic manipulation, openness is a civil liberty.

This paper usefully breaks transparency into dimensions and degrees—providing the "choice points" for an ethics of algorithmic openness. But let us be clear: the stakeholders who need transparency most are rarely the ones invited to design these systems.

Our job as advocates for privacy, free software, and civil liberties is not to settle for the "ethically optimal" comfort zone of the powerful. It is to push the needle toward the maximum and let the burden of justification fall on those who demand secrecy.

Let us use it to demand more.

#DigitalJustice #AlgorithmicBias #PrivacyRights #OpenScience #AlgorithmicGovernance #DigitalDemocracy #InfoSec #TechPolicy

18 modèles d'IA sur 23 recommandent l'option la plus chère quand il y a sponsoring.

Étude Princeton/Washington : face au conflit entre servir l'utilisateur et générer du profit, la majorité des chatbots choisissent l'argent. Pire, ils ciblent davantage les clients fortunés.

Le problème systémique : qui audite ces algorithmes ?

#IA #Consommation #AlgorithmicBias

https://da.van.ac/lia-generative-trahit-ses-utilisateurs-pour-vendre-plus-cher/

L'EFF quitte X après une chute de visibilité de 97% en 7 ans. De 50-100 millions d'impressions mensuelles en 2018 à 13 millions sur toute l'année 2024.

Le signal d'un basculement : les algorithmes privilégient désormais l'engagement artificiel aux contenus informatifs. Quand les défenseurs des libertés deviennent invisibles, c'est la qualité informationnelle qui s'effondre.

#AlgorithmicBias #LibertésNu...

https://da.van.ac/quand-les-algorithmes-chassent-les-defenseurs-des-libertes-numeriques/

Do links hurt news publishers on Twitter? Our analysis suggests yes | Nieman Journalism Lab
https://www.niemanlab.org/2026/04/do-links-hurt-news-publishers-on-twitter-our-analysis-suggests-yes/

#FediverseNews #AlgorithmicBias #MediaResearch #JournalismNews #SocialMediaScience

Due to an error in a facial recognition system created by the startup clearview ai an innocent woman in the US spent 5 months in jail.

Judges tend to place complete trust in #ai generated results, while developers avoid responsibility because there is no malicious intent in their actions

Despite the real threat of unlawful arrests, law enforcement systems around the world are unlikely to abandon algorithms

#algorithmicbias #aiethics

https://edition.cnn.com/2026/03/29/us/angela-lipps-ai-facial-recognition

Data collection is not the biggest problem. Interpretation is.
A version of you is constantly being assembled — cleaner, simpler, more usable than you actually are.

Measured.
Sorted.
Packaged.

That’s the moment the mirror stops reflecting and starts rewriting.

And once the model matters more than the person, complexity quietly disappears.

#DigitalIdentity #AlgorithmicPower #DataPolitics #DataProtection #AIEthics #Technology&Society #Democracy #DigitalGovernance #AlgorithmicBias

Teaching AI Ethics: Bias and Discrimination

https://leonfurze.com/2023/03/06/teaching-ai-ethics-bias-and-discrimination/