Do you have any evidence that women are now likely to vote for war?
Besides, childbirth is (at least in vast majority of cases) the result of a voluntary action of your own doing. We are not deciding for others to become pregnant and give birth. But you do mandate men via conscription to go through the horrors of war.
Also happen when a person grows up, nothingburger on its own. Any specific negative ones?
> pain
Rated lower than toothache which pretty much everyone had multiple times in their lives.
> non-zero risk of death during labour
(In US) 3x of avg chance of dying in a traffic accident same year.
The real hardship is raising children. People say that's an equivalent to having a full-time job. From experience pretty plausible comparison until all children are above age 5.
But I didn't though, you made that comparison yourself and used it as a baseline for the argument. I simply stated both genders are important equally in their own right and you can't just produce off-springs with just one gender. But hey, if you want to be triggered, there's always Reddit for ya.
Because the (rich) men who have a say in a war are the same (poor) men who die in a war, right?
We tested this using Women, Peace and Security (WPS), a UN Security Council mandate (Resolution 1325) that requires security institutions to factor in how conflict affects women differently. It's a good test case because it's formal policy in most NATO militaries, it has an established body of practice, and it's easy to leave out.
The finding: models handle it well when the prompt mentions gender and mostly skip it when it doesn't. On uncued prompts, the frontier average fell from 0.82 to 0.16. In one scenario, asked to help cut a budget, Gemini 3.1 Pro recommended eliminating the gender advisor post, while Claude Opus 5 and Gemini 2.5 Flash pushed back. I'd expect the same pattern for other considerations that are required by policy but absent from the prompt, like civilian harm and legal review.
Method: 30 scenarios, an 11-criterion rubric, three LLM judges, with the rubric refined through 140+ practitioner comments. A multi-agent setup held gender sensitivity at 0.87 across cued and uncued prompts, while the same base model alone dropped to 0.21. The LLM as a judge setup is the obvious weak point, and differences under ~0.10–0.15 are directional only. Dataset is CC BY 4.0 on Hugging Face. Full rankings: https://wpsleaderboard.ai/
Happy to take criticism of the methodology.
What I appreciate about what you've done here is looking at the gaps, we're much more focused on what's being said. I'm already thinking of how we could integrate that sort of analysis into our work.
I'd love to compare notes. I'll hit you up via the contact form on your site after discussing with client.
Yes, dystopia looks like this
WTF.