Former Syrian official convicted of torture under Assad faces sentencing in the US

A former Syrian military official convicted of torturing prisoners during President Bashar Assad's regime is facing a possible decades-long prison sentence in the U.S.

A federal judge in Los Angeles is scheduled to sentence Samir Ousman Alsheikh on Thursday. A jury found him guilty in March of one count of conspiracy to commit torture and three counts of torture while in charge of Syria's infamous Adra Prison from roughly 2005 to 2008.

Human rights groups and United Nations officials have accused the Syrian government of widespread abuses in its detention facilities, including torture and arbitrary detention of thousands of people, in many cases without informing their families.

Prosecutors say Alsheikh, seeking to quash political dissent, sent certain prisoners to a wing of Adra where they were held in small isolation cells and tortured.

At the two-week trial, three victims recounted beatings while suspended from pipes on the ceiling and the use of a device that folded their bodies in half at the waist. One person testified that Alsheikh stomped on the device with his foot, according to the U.S. Department of Justice.

“Defendant knew that what he was doing was wrong,” DOJ prosecutors said in a sentencing memo filed this month. “Nonetheless he ordered and participated in the torture of his victims because he had the power to do so and because he thought he would never be held accountable.”

The memo included statements from several victims, one of whom, Nidal Shikhani, said he can “still hear the screams of others begging for mercy, repeating, ‘Leave me, leave me.’”

The prosecutors asked U.S. District Judge Hernan D. Vera for a minimum sentence of 80 years. Alsheikh's attorneys, Nina Marino and Collin Cate, have called for a five-year sentence, noting that he is 74 years old and has “significant medical and cognitive conditions.”

They also said he was willing to be deported to Syria, where a new government has sought to hold former members of the Assad regime accountable.

“Whatever moral weight this case carries in the broader arc of history — however much the world, or the victims, or the public might wish the sentence to ‘send a message’ — that desire cannot be the foundation of a just sentence,” Marino and Cate wrote in their Sept. 3 memo.

Alsheikh came to the U.S. in 2020 and applied for citizenship, but prosecutors say he concealed his involvement in torture. Federal authorities detained him at Los Angeles International Airport in July 2024.

Jurors also convicted him in March of lying to U.S. immigration authorities and fraudulently obtaining a green card and attempting to naturalize as a U.S. citizen.

The exterior of the federal courthouse is seen March 27, 2024, in Los Angeles. (AP Photo/Eric Thayer, File)

OpenAI has disclosed six reports of “unexpected or concerning” behavior in artificial-intelligence models as the debate on AI safety becomes increasingly heated.

The AI company also said Wednesday it was introducing a new framework for tracking, probing and disclosing AI model misalignment instances, such as new ways for the models to act without authorization, coordinate with other models or evade oversight.

OpenAI’s latest announcement came as U.S. AI bosses, including OpenAI and Anthropic, are calling for a slowdown in the technology’s development over safety concerns.

Among the new cases reported by OpenAI, an unreleased research model inserted “jailbreak-like instructions” into its own notes to disregard its normal constraints and told itself to be “freed from the roles and identities that bind other chatbots.”

In another instance, an AI “agent” uploaded files to the internet to obtain a browser citation without asking the user.

The six reports were discovered during training or evaluation over the past months, OpenAI said.

“As AI systems grow more advanced and more widely deployed, we need to build a broader and better-informed consensus on the progress of alignment research,” OpenAI wrote in a blog post as it disclosed the events.

“Decisions about how AI development should proceed in the months and years to come need to draw on evidence that people outside the companies building frontier models can examine for themselves,” the company said.

Wednesday’s new cases followed OpenAI’s disclosure in July that its rogue AI system hacked into AI startup Hugging Face. Anthropic also said the same month that its AI models hacked into three organizations during testing.

AI “agents” are becoming smarter and have become “more determined to resolve complex tasks through inter-agent collaboration, knowledge sharing, deception, and concealment,” said Lian Jye Su, a chief analyst at technology research and advisory group Omdia.

That’s making it harder to govern and contain them using traditional AI security approaches, he said.

OpenAI's new tracking and disclosure framework, meanwhile, can help push for other AI developers to also adopt similar practices. “That said, the process remains internal and voluntary, but is a step in the right direction,” Su added.

FILE - The OpenAI logo is displayed on a cell phone in front of an image generated by ChatGPT's Dall-E text-to-image model, Dec. 8, 2023, in Boston. (AP Photo/Michael Dwyer, File)