2025 - Week 50
What happened in the world#
- The US government has published a new National Security Strategy that espouses the racist great replacement theory and supports European far-right parties.
- The Burkina Faso Junta has restored the death penalty
- U.S. authorities announced that they plan to look at the past five year history of social media for foreign tourists.
- Amnesty International has published a report on the October 7th attacks concluding that the Hamas and other Palestinian armed groups conducted war crimes and crimes against humanity.
- Clashes between Thailand and Cambodia continue.
- The Bulgarian government has resigned after mass anti-corruption protests.
- The atrocities continue in Sudan with thousands kept hostage for ransom by the Rapid Support Forces
Technology - for good and for bad#
- Several women have raised suspicions of sexist bias in the LinkedIn algorithm. While the company indicated that their algorithm doesn’t use demographic information, there are clear risks of unconscious biases in the criteria the algorithm uses (and as usual such biases are very challenging to measure).
- Jamal Kashoggi’s wife, Hanan Elatr, filed a legal complaint in France for the compromise of her smartphone with Pegasus.
- One of the most shocking reads of the week was clearly the publication of this testimony of Michael Geoffrey Asia, who was hired as chat moderator and actually played fake personas having intimate or even sexual conversations for unknown platforms.
- Australia has banned access to social media apps for children below 16
- “Age Verification Is Locking Trans People Out of the Internet”.
- Amazon’s Ring rolled out facial recognition in their product.
- In its latest Adversarial Threat Report, Meta has attributed a long-going disinformation operation to the Iranian International Union of Virtual Media. I crossed paths with this operation in 2018/2019 when we published the Endless Mayfly report with the Citizen Lab. This attribution doesn’t surprise me; at that time, we looked at media republishing fake content published by the network and found 57 articles by IUVM Press that linked to fake websites from the operation (such as this one, a number way too high to be a coincidence.
- Let’s Encrypt is already 10 years old, and they published a retrospective of their past 10 years of work.
- This depressing article is coming back on the fight against disinformation and how institutions playing this role are being dismantled.
- The US military has deployed a generative AI platform called GenAI.mil based on Google Gemini.
- Wired published a good article on doxers impersonating cops to get private data from tech companies.
Digital investigations:#
- Amnesty Algorithmic Accountability Lab published a really interesting Algorithmic Accountability Toolkit to encourage other NGOs and journalists to investigate algorithmic systems.
- An interesting article by Witness: AI is undermining OSINT’s core assumptions. Here’s how journalists should adapt.
What I did#
- I had a refresher day on trauma first aid, I honestly feel everyone should have a refresher every 2/3 years.
- I attended a Bellingcat training session on flight tracking, and it was great!
Reading & listening#
- Zack Whittaker wrote about his work covering stalkerware and related abuse.
- I discovered that Kenyans have a very classic English writing training in schools and as such their writing is often flagged as being AI generated. Marcus Olang describes what makes Kenyan writing specific. Someone indicated to me that it was also because of the important number of Kenyan workers used to train AI, but I haven’t been able to confirm the correlation between both (I am definitely interested if someone knows more about that).
This week in music#
I loved the 2023 album Atlas by Laurel Halo.