'Censorship': US urges overhaul to social media laws

Plans to give social media users the ability to opt out of algorithms on their feeds amounts to online censorship, the US government has warned.

In its response to the federal government's proposed digital duty of care laws, the Trump administration said it had serious concerns with the laws, urging for American-based platforms to be exempted.

"A framework that empowers governments to enforce vague definitions of 'harm' risks becoming a mechanism for viewpoint-based censorship," the US government said in a statement posted on the American embassy to Australia website.

"Platforms will be incentivised to over-moderate speech that should be protected to avoid regulatory penalties."

Under the digital duty of care, users on social media will have the option to remove an algorithm-driven feed and instead default to seeing content from people they follow directly.

Tech platforms would also be forced to protect under-18s from harmful posts such misogynistic content, pornography, posts that encourage eating disorders or content that glorifies crime.

The eSafety commission would also have powers to issue removal notices to social media platforms.

Prime Minister Anthony Albanese introduced the proposal in early September, releasing a draft for consultation.

But the US government said Australia needed to determine how the harm and risks would be determined and whether standard definitions would be needed.

The statement said calls to filter content would not be different than outright removal.

"The US government would likely consider such a measure as facilitating censorship," it said.

"We ask that Australia withdraw any plans to mandate that platforms algorithmically suppress content based on ambiguous, government-defined safety criteria."

The US had called for the federal government to explain its response to industry concerns, and how the definitions of what constituted harmful content would be determined.

The statement said US companies would be disproportionately impacted by the measures.