Bot-like social media accounts have spread "disinformation and hate" in tens of thousands of posts viewed an estimated 150 million times during the UK general election campaign, a watchdog investigation revealed Tuesday.
Global Witness found 10 suspected bot profiles on X (formerly Twitter) have shared more than 60,000 messages containing conspiracy theories and violent hate speech, including Islamophobia, anti-Semitism, homophobia and transphobia, before Thursday's vote.
The NGO assessed that the accounts, which also promoted praise for Russian President Vladimir Putin, may have had "an oversize influence" given their prolific output.
Bot profiles are run by computers but have been programmed to look like human-run accounts.
The findings follow warnings that the integrity of key elections around the world this year could be threatened by rapid advancements in cyber-tech, particularly AI, and increasing friction between major nations.
UK Deputy Prime Minister Oliver Dowden warned on Sunday that hostile actors such as Russia were seeking to influence the country's democratic process after reports emerged of Facebook pages spreading identical pro-Kremlin talking points.
Ava Lee, campaign leader at Global Witness, urged X and other social media companies to "clean up their platforms and put our democracies before profit".
"Political discussion online is often toxic -- we all know that. But when we go on social media, we believe we're seeing what real people think," she said.
"When that's not true, when the conversation may have been influenced by someone who has paid for bots to spread division or to get a particular party into power, our democracy is in jeopardy."
- Overtly party political -
Global Witness researchers identified the sample of suspected bot accounts by searching for posts and hashtags on climate change and migration -- two hot button issues often subject to disinformation.
The team then identified and tracked at least 10 profiles pumping out scores or hundreds of posts daily on these issues in the weeks after the UK election date announcement on May 22.
Eight were overtly party political, clearly aligning themselves for or against a particular party by using its logos as their profile picture or regularly re-posting its content or hashtags, according to the NGO.
It did not find evidence that any UK party was paying for, using or promoting the bots as part of their campaigns.
The content shared by the suspect accounts was overwhelmingly extreme, Global Witness said.
"Some spread anti-Semitism and transphobia. Some state that climate change is a 'hoax', that vaccines have created a 'genocide'," it added.
One of the 10 profiles was recently deleted, but the other nine remained active as Global Witness released its findings.
It urged X -- known as Twitter until US billionaire Elon Musk bought the platform in 2022 and rebranded it the following year -- to investigate the accounts, which likely violate its own policies.
Global Witness also demanded X invest more in "protecting our democratic debate from manipulation".
X did not respond to requests for comment.