🔍 THE BOTTOM LINE
AI isn’t just changing what we do — it’s changing who gets to do it, how we’re judged, and whether we can trust the systems making decisions about us.
1. 👩 The AI Gender Gap Isn’t About Access — It’s About Penalty
The story: Multiple studies confirm women are 20% less likely to use AI than men, but the reason isn’t just access. When women use AI at work, they’re perceived as less competent. Men get praised for “innovation,” women get flagged for “cheating.”
Key facts:
- Harvard Business School study: women 20% less likely to use AI globally
- SSRN paper (2026): women using AI rated as less competent than male colleagues doing identical work
- Lean In research: men 27% more likely to be praised for AI use
- Internet NZ: 51% of women concerned about AI vs 42% of men
Why it matters: This isn’t a “pipeline problem” you fix with training programs. It’s a bias problem. Same behavior, different judgment based on gender. That doesn’t get solved with upskilling.
Our take: If your AI adoption strategy doesn’t account for this, you’re not doing DEI — you’re doing optics. The penalty is real, and it’s measurable.
2. 🤖 World Models: AI That Understands Physics, Not Just Patterns
The story: 2026 is the breakthrough year for AI “world models” — systems that simulate cause-and-effect, not just correlation. This is what enables robots to interact with the physical world without breaking things.
Key facts:
- World models enable real-time physics simulation for training embodied AI
- Interactive Genie-like systems emerging for robotics training
- Shift from “predict next token” to “simulate next state”
- Continual learning prototypes moving from research to deployment
Why it matters: Current AI is brilliant at “what usually comes next” and useless at “what happens if I do this.” World models fix that. It’s the difference between a chatbot and a robot that can make you coffee without flooding the kitchen.
Our take: This is quietly the most important AI development of 2026. Robotics + world models = AI that can interact with reality, not just describe it.
3. 📉 The AI Adoption Gap: Who Gets Left Behind
The story: Internet NZ data shows 69% of men use AI weekly vs 51% of women. That’s not just a usage gap — it’s a career trajectory gap. In 2 years, AI fluency will be baseline, not differentiating.
Key facts:
- 69% of men use AI weekly, 51% of women (Internet NZ)
- 51% of women concerned about AI vs 42% of men
- Women concentrated in frontline roles with less AI exposure
- When women do use AI, they’re penalized (see #1)
Why it matters: This is a compounding disadvantage. Less exposure → less fluency → less confidence → more penalty → less use. That cycle doesn’t fix itself.
Our take: Employers: if your AI training program doesn’t account for this, you’re not solving the problem — you’re widening it. Design for the people who need it most, not the people who already get it.
4. 🌐 AI Harassment and Deepfakes: The Election Threat
The story: Jo Cribb’s analysis highlights automated harassment campaigns, explicit deepfakes, and sextortion as growing threats — particularly for women in public life. This isn’t hypothetical; it’s already shaping NZ’s election environment.
Key facts:
- A third of NZ women report experiencing online abuse
- Deepfakes of politicians announcing fabricated policies already circulating
- AI-generated images deployed as political propaganda
- Data-driven voter targeting at scale
Why it matters: The “AI avalanche” isn’t coming — it’s here. By the time you see it, you’re already buried. The question is whether we build fences before or after the avalanche hits.
Our take: Transparency about where AI is used in public services, independent auditing for bias, and stronger regulation to prevent online harm. The best time to build the fence was before the avalanche. Second-best time is now.
🔍 THE BOTTOM LINE (reprise): AI is reshaping who gets to participate, how we judge competence, and what we owe each other in public discourse. The technology is moving faster than the ethics, the governance, and the infrastructure. That gap is where the problems live.
Sources:
- Harvard Business School — Gender gap in AI adoption study (2026)
- SSRN — “AI Use and Perceived Competence: Gender Differences” (2026)
- Lean In — “AI Women Gender Gap” research (2026)
- Internet NZ — AI understanding and concern survey (2026)
- NextBigFuture — “2026 is Breakthrough Year for Reliable AI World Models” (10 Apr 2026)
- Newsroom — “Immediate action needed to stop AI avalanche burying election debate” (Jo Cribb, 6 May 2026)