X has refused to take down dozens of social media posts reported as “hate, abuse or harassment” in which prominent UK politicians, including Kemi Badenoch, have been racially abused.

In May, researchers from the social inclusion thinktank British Future reported 30 posts from this year in which the Conservative party leader was called the N-word. In each case the researchers used the platform’s “hate, abuse or harassment” reporting option. X refused to act in the majority of cases, despite repeated requests.

The Guardian understands X routinely takes action only when posts are reported to it as illegal under the UK’s Online Safety Act. In those cases, it restricts visibility in the UK, leaving the post unrestricted in other jurisdictions.

Avaes Mohammad, a researcher with British Future’s British South Asian Bridgers cohesion project, said: “Many people will intuitively report racist posts as ‘hate, abuse or harassment’ – but it seems X doesn’t consider it as hate. Our research found that it was only when a post is described as illegal, which they can’t dispute, that they will consider taking it down.”

X eventually restricted visibility for just two of the 30 messages in which Badenoch was called the N-word – and only after Sunder Katwala, the director of British Future, emailed the platform saying he should not be expected to fill out an “onerous” form reporting each one as illegal.

Elon Musk has been using X to amplify far-right narratives since he bought the platform in 2022. Keir Starmer earlier this month accused Musk of trying to “whip up division” with his tweets about the murder of the student Henry Nowak in Southampton.

The Online Safety Act, which is being phased in by the communications regulator, Ofcom, places obligations on platforms to remove illegal content, which can include racially or religiously aggravated offending.

On 15 May – the same day Ofcom announced X had made a voluntary commitment to remove illegal content within a 48-hour window – British Future reported 33 uses of the P-word targeting UK public figures.

Reform UK’s Zia Yusuf was told in one tweet, which used the P-word, to “sod off” to Pakistan. Others who were racially abused included the home secretary, Shabana Mahmood; the former SNP leader, Humza Yousaf; the Reform UK London mayoral candidate, Laila Cunningham; the Advance UK founder, Ben Habib, and the founder of Your Party, Zarah Sultana.

Forty-eight hours later, none of the 33 posts reported had been removed.

When British Future emailed X to ask why, the platform advised it to report the posts using the “UK illegal content” option from the “report post” menu, adding: “Reports made through other channels are handled under our general policies and do not count toward the specific expedited review targets under the Online Safety Act.”

X eventually restricted 20 of the tweets within the UK after being contacted by Ofcom, it is understood, but took no action against the rest. Among messages X left untouched was a picture of a noose tweeted to Sultana, saying “fuck off” and using the P-word.

On 26 May, British Future reported another nine posts that used the P-word against British Asian public figures. In six, X failed to act because the posts were reported as “hate, abuse or harassment”. It was only when posts were reported as illegal, in the remaining three cases, that it acted.

None of the account holders behind any of the offensive tweets have been suspended, including one who used the N-word 45 times in a week.

British Future says the platform’s approach means racists can operate with impunity on X.

“The accounts we are reporting are often replete with mentions of Reform and Restore,” Mohammad added. “In May, the N-word racist slur was used about Kemi Badenoch an average of once a day. But on 2 June, after Badenoch responded to Nigel Farage’s speech about Henry Nowak, there were 16 examples in a day – pointing to how much far-right political sentiment is driving the hate on X.”

Ofcom said it condemned racism “in all its forms”, and its job was to ensure sites had “appropriate measures in place to comply with their duties” rather than telling them what to take down.

The regulator said X had committed “to reviewing and assessing UK suspected illegal terrorist and hate content reported through its dedicated UK illegal content reporting tool” within an average of 24 hours, adding: “We’ll be carrying out quarterly reviews of X’s performance.”

X did not respond to requests for comment.