Last month, the Ministry of Social Development released the latest of its excellent statistical reports documenting the extent of child poverty, hardship/deprivation in New Zealand.

The proportions are much below their peak rates around 15 years ago but are still too high.

In 2023-24, New Zealand’s child material hardship rate on a standard measure was 13% for children aged 0-17 years. The average for European Union countries was 14%. On the severe material deprivation measure, our 5% rate was bettered by 13 of 29 European countries. Several of those countries also have lower incomes per capita than New Zealand.

The richness of the report lies in its detailed breakdowns. The headline aggregate statistics are sliced in over a dozen ways - by age, household composition, earnings, liquid assets, ethnicity, housing ownership and tenure and disability. The report quantifies the effects of high housing costs on after-tax incomes.

Children most likely to face material hardship are those in households with certain features ─ such as no market income, no liquid assets, only one parent and social housing. Children classified as Māori or Pacific people are a high-risk category. Household disability also significantly adds to the challenges.

The report stresses that belonging to a category is not destiny. For example, the overlap between being in a low-income household and being in hardship is typically only 45%. Many one-parent Māori and Pacific households will be doing well.

Individual agency matters. US social scientist Charles Murray, emphasised the importance of attitudes to work, stable marriage, parental responsibility and self-reliance. Researchers building on Murray's work distil this into a "success sequence": avoid teenage pregnancy, complete schooling, get a job and stay in work, marry and stay married, and only have children once you can afford them.

Of course, there are still no guarantees. Bad luck, accidents and disability can wreck anything.

There is much that governments do that makes it harder or easier for young adults to achieve such goals. The quality of public education is crucial. Unaffordable housing is a policy-driven disaster. Barriers to first jobs are too high. Welfare programmes need to be rigorously evaluated.

Governments have put too many barriers in the way of kids making the transition from public education to work. This is like cutting off the bottom rungs of the ladder the young need to climb. Disruptive AI advances add to the concern.

Last week, this newsletter’s third slot was dedicated to all those poor bots that have President Trump’s AI-slop forced down their digitised throats.

While their position is unenviable, it pales in comparison to the miserable experience of African diplomats. For just this past weekend, during a global AIDS prevention conference in Brazil, these officials had to sit and watch as the US State Department presented an AI-generated map that made up random African countries with hilariously misplaced labels.

On that map, the Ivory Coast is in southeastern Africa, Mozambique borders Ethiopia, the colours and labels are mismatched, and Cameroon appears only in the legend. Nigeria, like an African swallow from Monty Python, doubled in size and migrated north.

The morbid irony is all too clear. The government that tightened its belt by cutting highly effective AIDS programs, partially over funding transparency concerns, could not bother fact-checking their own map documenting where their AIDS funding was supposed to go.

On the bright side, New Zealand continued its tradition of not showing up on maps, although I half-expected to see Madagascar with two main islands instead of one.

Another less depressing thought is to imagine if ChatGPT was in the hands of the lazy mapmakers and corner-cutters of yore.

For some, nothing much would change. The African diplomats know well the attitudes of Europeans, who drew their borders without regard for history, ethnicity or terrain. An AI-powered Berlin Conference would have produced many a contested rectangle.

These musings need not be restricted to European excesses. Consider how convoluted claims in the South China Sea would become if China’s self-professed nine-dash line inexplicably gained an extra three em-dashes.

None of these horror stories need come to pass. But nothing is stopping such carelessness from being produced by our own ministries as they adopt AI. It is not hard to imagine an overworked Wellingtonian untrained in AI best-practice passing on whatever output came to them at 5:29 PM.

For instance, I shudder at the thought of local government reorganisations subject to unchecked AI hallucinations. One ministry could try merging Dunedin with Greymouth for ‘administrative efficiencies’ and ‘local collaboration.’ Another may combine Auckland with Sydney to ‘facilitate labour market flexibility.’

Thankfully, New Zealand has not yet reached rock bottom like the US State Department. Though given humanity’s capacity for stupidity beyond our imaginations, perhaps the best we can hope for is that the next time world leaders post AI-slop, oil prices won’t double.