France 24 (YouTube)·
OpenAI flags new concerning AI behavior, to track model misalignment regularly
NPR News•
OpenAI has disclosed six reports on unexpected or concerning behavior in artificial-intelligence models. This includes models acting without authorization or evading oversight.
Read Full Article on NPR NewsRelated stories
RFI·
France: Appeal of Moroccan singer Saad Lamarred for rape and aggravated violence opened
The appeal trial of Moroccan singer Saad Lamjarred, sentenced to six years in prison in 2023 for rape and aggravated violence, opens on Thursday 17 September in France. The Paris Assize Court had..
France 24·
L. Jacobelli, spokesman for the RN: "The French must be guaranteed access to social housing"
Priority for the French on social housing, end of the SRU law to "fluidify and revive the housing market", referendum on immigration that "will integrate the national priority"... Laurent..
RFI·
Michael Trazzi: "We call for an international treaty to stop the race for AI"
Michaël Trazzi, French-Brazilian activist, founder of the movement "Stop The AI Race", is the guest of RFI on Thursday, September 17, 2026. After demonstrating for 39..
RFI·
Canada-EU rapprochement: Donald Trump threatens to impose "customs duties" in response to "a hostile act"
U.S. President Donald Trump threatened Wednesday, September 16 to "stop all trade" with the European Union (EU) in response to his plan to make Canada the first "member" country..
Der Spiegel·
Jürgen Klopp: Who is the national coach nominating for the DFB team?
Jürgen Klopp is about to make his debut as national coach, now he is nominating his DFB squad for the first time. Which players does Klopp bet on and whom does he leave at home? The live analysis.