IT’S easy to start feeling detached from reality when listening to politicians speak. Not just when they give slippery answers to straightforward questions, but also when they loftily claim to speak on behalf of the nation, presuming to tell us what we’re concerned about and how they have the answers.

To these perennial problems, a new one has been added: it’s getting harder and harder to know whether what we are seeing or hearing from public figures is actually real. Increasingly, a photograph or video of a politician isn’t enough. Authentication is required too.

When Keir Starmer gave an interview to Sky News on February 22, responding to questions about the House of Commons debacle on the SNP’s Opposition Day debate the day before, he made reference to “having spoken to the president of Israel”, among others, before crafting Labour’s Gaza amendment.

Sky published a video of the interview on its website, under the heading “Starmer ‘categorically’ denies threatening Speaker over Gaza vote” – but a post on X/Twitter including a clip of it was deleted after users raised questions over the referenced meeting with Isaac Herzog.

The National reported on the deletion of the post, and our coverage is the first Google result under “news” for the search “Starmer met Israeli president”.

READ MORE: Jeremy Corbyn announces legal action against Nigel Farage

All of the other stories on page one of the search results are from January 6 or earlier – long before Starmer made the statement.

While the video has, unsurprisingly, circulated on social media, most savvy consumers of digital media know better than to assume an isolated clip hasn’t been dubbed, or even completely faked using artificial intelligence (AI). If there’s little or no mainstream media verification that the statement was really made, is it sensible to believe it?

The BBC reported yesterday that supporters of Donald Trump have been targeted with AI-generated images portraying the US presidential candidate surrounded by smiling black people.

One of these, shared on the BBC’s website with a large “FALSE” across the image, was created by staff on a conservative radio show, and shared on Facebook to accompany an article about black Trump supporters.

The show’s host, Mark Kaye, was unrepentant about these actions, telling a BBC reporter: “If anybody’s voting one way or another because of one photo they see on a Facebook page, that’s a problem with that person, not with the post itself.”

The National: Keir Starmer

The image has some of the hallmarks of an AI fake: one of the people depicted is missing part of a finger, and Trump’s own digits are both webbed and oddly smooth for a man in his 70s. But in some respects this fiction is less suspicious than some of the genuine recordings of both Trump and the current US president Joe Biden speaking.

Addressing an audience in Virginia at the weekend, Trump said: “Putin, you know, has so little respect for Obama that he’s starting to throw around the nuclear word ... we have a fool, a fool as a president.”

Journalists were quick to report that it was the third time in less than six months that he had muddled the past and present presidents.

Biden, who at 81 is four years older than his rival, has become well-known for verbal gaffes, the latest of which saw him repeatedly refer to sending supplies to Ukraine during a conversation with the Italian PM that was actually about Gaza.

Against a backdrop of well-documented fumbles by both candidates – and during discussion of major global crises, too – it would not be difficult to introduce doctored clips to the mix and have voters believe them to be real.

READ MORE: 'Back off': MPs slate plans to ban them engaging with pro-Palestine groups

The technology available now is much more sophisticated than it was when Trump ran against Hillary Clinton, when one popular activist tactic was to splice together existing footage of the latter joking around or making sudden movements in an attempt to portray her as seriously ill or mentally unstable.

The best defence against the creep of AI is well-resourced media that will report without fear or favour, and our own natural intelligence, which we can deploy to sniff out images from the uncanny valley or language that doesn’t sound quite right.

It’s concerning, then, that a controversial statement from an opposition leader seems to have ended up buried rather than interrogated, and also that the man currently running the UK seems happy to issue urgent statements that, while increasing tensions and worry, don’t entirely make sense.

In his evening statement outside Downing Street on Friday, Rishi Sunak earnestly warned about efforts to “turn impressionistic minds against their own society”. Is the Prime Minister particularly worried about latter-day Monets becoming radicalised, or was his speech generated by the same technology that spat out the scripts for Willy’s Chocolate Experience?

As voters, we need to keep our wits about us and question what we are being shown (or not) and why. But if politicians keep tripping up, that job will be harder than ever.