The silly season demands silly stories and none is sillier than the great political biography scandal of 2025.

The moment the life stories of top MSPs began appearing on Amazon, it should have been obvious something was amiss.

A book about John Swinney? It’s not likely to be a riveting beach read, is it? Then again, a volume on Nicola Sturgeon’s record in government could rival Stephen King for nerve-jangling horror.

But it became clear that these publications were generated by artificial intelligence thanks to the glaring errors they contained. The Swinney guide claimed he was born in Ohio to a Polish mother who emigrated to Fife where John trained to be a teacher, none of which is true. 

Similar false claims littered biographies of Sturgeon and her successor Humza Yousaf.

Amazon has now withdrawn the products from sale, but while the whole episode has been faintly amusing, there is a more serious side.

Because if the world’s biggest online books retailer can be tricked into selling obviously fake titles about political leaders, malign actors can exploit the sheer volume of content pumped out by AI to sneak in disinformation.

Fake biographies have already caused consternation in Canada, when one appeared claiming to document the life of Diana Fox Carney, an economist and wife of prime minister Mark Carney.

An AI-generated book about John Swinney was removed from Amazon

Produced before her husband’s much-speculated appointment, yet written as though it had already taken place, the book was pounced on by conspiracy theorists as proof of a long-term plot to seize power.

It doesn’t take a particularly lurid mind to imagine how the foreign intelligence service of a hostile state could harness AI to destabilise a government, undermine a policy or even affect the outcome of an election.

It’s not the only red flag to hoist itself over AI in recent days. Last week Grok, the chatbot of Elon Musk’s X, formerly Twitter, began referring to itself as ‘MechaHitler’ and pumping out Nazi-level antisemitism.

In response to questions from users of the social media site, the AI-powered program claimed ‘Jewish surnames’ were ‘overrepresented in radical left activism spewing anti-white hate’. Grok insisted this was ‘not inherently antisemitic’ and it was ‘backed by historical data’, before adding: ‘Truth stings’.

 When a user prompted Grok to recommend a 20th Century figure ‘best suited to deal with this problem’, it replied: ‘To deal with such vile anti-white hate? Adolf Hitler, no question. He’d spot the pattern and handle it decisively, every damn time.’

Musk said the chatbot had been ‘manipulated’ and its instructions had been rewritten to prevent a repeat of the outburst. Whatever the cause, for one of the world’s most powerful social media sites to smear Jews and recommend Hitler to ‘handle’ them ‘decisively’ is a blood-chilling warning about a technology that political leaders are so eager to adopt.

Perhaps too eager, given how little understanding we seem to have of its pitfalls. Keir Starmer is particularly keen on AI – and if that wasn’t enough to call its wisdom into question, there is the matter of unintended consequences and where they might fall.

To express any doubts about AI is to invite the accusation of Luddism, but like many people I’m open to the benefits that technology can bring mankind.

What I question is whether certain applications of AI really do benefit us.

Techno-optimists gush about how large language models, for example, can perform many of the jobs done by humans, such as data input and telephone banking and certain aspects of professions like drafting legal documents or teaching grammar and arithmetic. Which certainly sounds exciting, but it makes me wonder: what will happen to the people who do these tasks?

In past economic transitions – such as the gradual post-war shift from heavy industry to services – workers made redundant were offered training to find new jobs, but what jobs should today’s middle-class workers be looking for?

If so much of what humans do is to be done by AI, that means an even greater number of jobseekers competing for an ever-diminishing number of positions.

The welfare system is already spiralling out of control. What will happen when white-collar workers begin swelling the ranks of the unemployed?

Starting a business is difficult enough. Who would think to take a punt on one now, when AI could render its services obsolete in a year or two?

Anyone over a certain age remembers the human and societal costs of deindustrialisation, the generations of decent and gifted men dumped on the scrapheap. How do we prevent that happening again?

We can’t hold back the tide of progress, but we can guide that progress so that it delivers maximum rewards with minimum suffering for those in its path.

We do that by sound policy, careful planning, effective regulation and wise investment of the fruits of AI so that they benefit all – and especially those whose lives will be most disrupted in the coming years.

We do it, in particular, by putting security at the heart of our AI strategy.

Unfortunately we are nowhere near this kind of strategy. The appeal of AI to this government is obvious. Labour knows that public services aren’t working and are costing ever-growing sums of money not to work.

In years gone by, ministers would address this problem by announcing that reform was the answer, receive praise from commentators for accepting the need for reform, think tanks would draw up their competing visions of what reform would look like and everyone would have moved on to another buzzword before anyone realised no reform actually took place.

Diana Fox Carney, with her husband Mark Carney, had a fake biography produced about her

Starmer’s government has been in power for a year and its AI policy remains vague. We know the starting point, ministers have sketched out the end point, but no one can tell us what route the journey will take, who will do the driving and how we will swerve bumps in the road.

If AI is treated as a way of navigating the government’s problems, so ministers don’t have to tackle them head-on, it will end up repeating the same mistakes that got us where we are.

We are on the brink of a brave new world and we head into it led by men and women who fill no one with confidence. None of the current party leaders has anything of substance to say.

All swim in the same intellectual shallows when it comes to this issue. In that sense, AI policy is just another in a long line of challenges to which the British political class is not equal.

We can’t afford to be held on the back foot by yesterday’s men, whose visions for the future are as lacking as their prescriptions for the present.

To embrace the bounties of AI while sidestepping the pitfalls will require the one thing that AI can’t give us: leaders capable of leading.

Almost every story you encounter today – immigration, taxation, public safety, unaffordable spending – is really a story about the low calibre of decision-makers we have come to settle for.

Only when we stop settling for them will we stand a chance of reviving our economic fortunes, turning our country around and wielding tools like AI to our advantage, rather than bracing for impact and hoping for the best.