Amnesty Tech

Amnesty International - Tech.
We are a global collective of advocates, campaigners, hackers, researchers & technologists defending our rights in a digital age.

[Amnesty Tech @AmnestyTech]

🚨NEW: @Amnesty briefing documents how mass data pipelines of gen AI systems are rooted in invasions of privacy by design.
https://t.co/z1UnJlko09

[Amnesty Tech @AmnestyTech]

If you are in New York for #UNGA, please do come to our side event with our partner @humanusofficial. We'll be discussing how to improve access to remedy for Big Tech harms in conflict-affected settings, and the lessons we need to take forward in the era of AI.

[Amnesty Tech @AmnestyTech]

The UK & France must prioritize human rights as they lead the Pall Mall Process to develop Industry Guidelines on Commercial Cyber Intrusion Capabilities, including spyware to ensure these tools are not used to harm people, communities & institutions.

https://t.co/waEPDkjjI3

[Amnesty Tech @AmnestyTech]

Digital technologies are being used to gatekeep asylum systems.

"Asylum Is Unavailable," a new book by @StephenDamianos exposes how the Greek Asylum Service used a skype line between 2014 and 2022 as part of a digital push back against asylum seekers.

https://t.co/KD3fExkdaD

[Amnesty Tech @AmnestyTech]

Artificial intelligence dominated the news this week.

@amnesty is calling for an international, rights-based & multi-stakeholder regulatory framework, not one decided by just a few powerful states & companies.

States must pause all development & deployment of harmful AI tech.

[Amnesty Tech @AmnestyTech]

.@EU_Commission must prioritise safety-by-design for all users as opposed to implementing age assurance methods which increase personal data collection when there are already concerns that technical solutions currently being tested do not sufficiently protect online privacy.

[Amnesty Tech @AmnestyTech]

The @EU_Commission has proposed in #SOTEU a social media ban for children under 13, “mini accounts” for 13- to 14-year-olds with limited features, a one-hour time limit & an obligation for tech platforms to provide 15 to 18 year-olds with safe-by-design accounts.
@GettyImages

[Amnesty Tech @AmnestyTech]

For example, in abusive family relationships, as well as potential discriminatory effects on children in the care of guardians other than their parents, practical concerns around some parents’ ability to take up this task & privacy risks associated with the added data trail.

[Amnesty International @amnesty]

As the demand for generative AI has snowballed, the scale of the development of these technologies has spun out of control.

Complex supply chains are required to power these systems, presenting us with a trail of human rights risks at every step of their production.

These design choices must be challenged.

They are not inevitable.

[Amnesty Tech @AmnestyTech]

We’ve joined 140+ organisations in calling on @EU_Commission President @vonderleyen to ensure tech companies prove their products are safe & age-appropriate before they reach children.

Design for safety not exploitation.

#SOTEU

https://t.co/FxrzJnYFI4

[Amnesty Tech @AmnestyTech]

🚨 New @amnesty research shows how @Meta failed to effectively moderate anti-LGBTI content on @facebook, the main social media platform in Hungary, even when reported. Head of Big Tech Accountability Team @amnesty Alia Al Ghussain, explains how this affects the LGBTI community.

[Amnesty Tech @AmnestyTech]

The EU must robustly enforce the Digital Services Act (DSA). Hungary’s new government must ensure the DSA is implemented & that the country’s Digital Services Coordinator (DSC) is independent & properly resourced.

https://t.co/6pbUNeFODc

[Amnesty Tech @AmnestyTech]

In January 2025, @Meta changed its moderation policies, weakening safeguards for high‑risk groups. Our latest research shows how @Meta failed to adequately assess the risk to LGBTI people in Hungary and address harmful rhetoric against them.

https://t.co/rNfMfFDdP3

[Amnesty Tech @AmnestyTech]

A social listening exercise done as part of this research found Facebook to have a particularly hostile environment towards LGBTI-related discussions. In some cases, LGBTI people were referred to as threats to children, families & traditional values.

[Amnesty Tech @AmnestyTech]

Meta reached an $18 billion settlement with US states over child safety & committed to make design changes to its social media apps.

This is a turning point not just for Meta but all social media. They must fix their harmful features to protect children.

https://t.co/4EYZDE3Yo9

[Amnesty Tech @AmnestyTech]

Ahead of Brazil’s presidential elections, we carried out a social media experiment to test @facebook's moderation of ads containing election-related disinformation. We found moderation weaknesses.

@Amnesty Head of Big Tech Accountability Alia Al Ghussain explains the findings.

[Amnesty Tech @AmnestyTech]

.@Amnesty is calling on @Meta to ensure adequate Brazilian Portuguese language resourcing for content moderation including ad approvals and to immediately strengthen its capacity to identify and respond to election-related harms.

[Amnesty Tech @AmnestyTech]

Ahead of Brazil’s presidential election @amnesty submitted @facebook ads, to test @Meta's moderation system but with no intention of publishing them. Over half of the disinformation ads were accepted, while the rest were rejected but not directly for containing disinformation.

[Amnesty Tech @AmnestyTech]

The investigation indicates that @facebook's moderation systems may be failing to reliably detect election-related disinformation content. This raises serious concerns about whether @facebook's election safeguards are being consistently applied.