Experts in learning, not AI. Research, analysis, and perspectives on AI capability.https://ablative.com.au/https://ablative.com.au/blog/the-lighter-load/https://ablative.com.au/blog/the-lighter-load/The research that proves you're already doing enough. Four non-burdensome, research-backed findings for parents who are carrying too much.Sat, 01 Aug 2026 00:00:00 GMTAnnabel Grayhttps://ablative.com.au/blog/three-things-every-child-needs/https://ablative.com.au/blog/three-things-every-child-needs/And 90 years of research that tells us the order they need them in.Wed, 01 Jul 2026 00:00:00 GMTAnnabel Grayhttps://ablative.com.au/blog/ai-decisions-children/https://ablative.com.au/blog/ai-decisions-children/Australia's new Support Needs Assessment tool uses computer-generated systems to determine children's access to disability support services — without transparent explanations of how decisions are made.Wed, 01 Jul 2026 00:00:00 GMTAnnabel Grayhttps://ablative.com.au/blog/suspension-wrong-problem/https://ablative.com.au/blog/suspension-wrong-problem/Reading that 150 students a day are suspended from Victorian schools makes the problem impossible to ignore. Suspension treats the symptom as if it were the cause.Mon, 01 Jun 2026 00:00:00 GMTAnnabel Grayhttps://ablative.com.au/blog/how-you-treat-ai/https://ablative.com.au/blog/how-you-treat-ai/At Ablative our AI agents have names. They all have roles, responsibilities, and things they're good at. This may sound crazy to some people — but I can almost guarantee in a few years it won't.Mon, 01 Jun 2026 00:00:00 GMTAnnabel Grayhttps://ablative.com.au/blog/stretch-points/https://ablative.com.au/blog/stretch-points/The words we use mean things, and as such — have consequences. Research has proven that the framing we use for children has an impact on the way they are treated and, as a result, their outcomes.Mon, 01 Jun 2026 00:00:00 GMTAnnabel Grayhttps://ablative.com.au/blog/harness-asymmetry/https://ablative.com.au/blog/harness-asymmetry/Same model, different harness, fifteen points of benchmark. Opus 4.7 launched yesterday and landed twelve points below Opus 4.6-on-a-better-harness on the same benchmark. The AI industry has learned that the scaffolding around a model does most of the work. It has not yet noticed the same is true of the humans.Wed, 01 Apr 2026 00:00:00 GMTAblative Researchhttps://ablative.com.au/blog/human-harness/https://ablative.com.au/blog/human-harness/If a harness around the human is what's missing, the engineering question is what one actually looks like. Item response theory has been building exactly this kind of scaffolding since the 1970s. The mathematics doesn't care whether the thing being assessed is a nurse, a child, or Claude Opus 4.7.Wed, 01 Apr 2026 00:00:00 GMTAblative Researchhttps://ablative.com.au/blog/embed-problem/https://ablative.com.au/blog/embed-problem/In 1885 a German psychologist locked himself in a room with a list of nonsense syllables and proved that people forget things. A hundred and forty years later, HR departments are still funding the one-day workshop as though the memo had mysteriously failed to arrive.Wed, 01 Apr 2026 00:00:00 GMTAblative Researchhttps://ablative.com.au/blog/trust-paradox/https://ablative.com.au/blog/trust-paradox/Forty-nine percent of Australian workers use AI at work in ways that violate their employer's policies — the highest rate in a forty-seven country study. It is, as far as anyone has measured, our most consistently-executed national AI strategy. Nobody approved it. Everyone's following it.Wed, 01 Apr 2026 00:00:00 GMTAblative Researchhttps://ablative.com.au/blog/headcount-expectation-gap/https://ablative.com.au/blog/headcount-expectation-gap/The RBA asked 105 Australian employers about AI. Half quietly expect to shrink their headcount. The Tech Council asked the workforce. Ninety-three percent expect to be augmented, not replaced. Both parties are planning for the future with great confidence, and they have not compared notes.Wed, 01 Apr 2026 00:00:00 GMTAblative Researchhttps://ablative.com.au/blog/australia-commitment-gap/https://ablative.com.au/blog/australia-commitment-gap/Eighty-four percent of global enterprise leaders plan to raise AI investment next year. Sixty-five percent of Australian ones do. The consultants call this a commitment gap. Measured against the 5.5% of firms globally that actually capture meaningful return, it might just be called reading the data.Wed, 01 Apr 2026 00:00:00 GMTAblative Researchhttps://ablative.com.au/blog/evidence-supply-chain/https://ablative.com.au/blog/evidence-supply-chain/We mapped who produces AI adoption research, who cites whom, and where the conflicts of interest are. The evidence base is dominated by its own customers.Wed, 01 Apr 2026 00:00:00 GMTAblative Researchhttps://ablative.com.au/blog/forty-eight-cells/https://ablative.com.au/blog/forty-eight-cells/Four dimensions, three domains, four modes. An interactive exploration of the framework that measures what AI maturity models miss.Wed, 01 Apr 2026 00:00:00 GMTAblative Researchhttps://ablative.com.au/blog/adoption-anti-patterns/https://ablative.com.au/blog/adoption-anti-patterns/We read 39 AI adoption studies. The data tells a consistent story — just not the one most reports are trying to tell. Here are six recurring failure modes, with the numbers behind each one.Wed, 01 Apr 2026 00:00:00 GMTAblative Researchhttps://ablative.com.au/blog/capability-gap/https://ablative.com.au/blog/capability-gap/Most organisations measure AI adoption by tool deployment. We measure it by what people can actually do. The difference is everything.Sun, 01 Mar 2026 00:00:00 GMTAblative Research