Así es. Ese "no tenemos tiempo para hacer estudios" me parece increíble

Desde los años 2000 hay múltiples ejemplos de cómo la gran industria tech no ha frenado cuando "las cosas no iban bien" con sus productos y servicios. Más bien al revés. Confiar en que ahora lo hagan me parece absurdo. Y sobre todo, peligroso.

Y continúa: "Es complicado que se pare o se ralentice. Pero espero que si tenemos suficientes datos de que una cosa no va bien podamos frenar un poco."

"Yo pienso como científico pero también como padre, pienso el mundo que vendrá y cómo cambiará (...). Pienso profundamente en las repercusiones, en si es bueno que los jóvenes tengan acceso a chatbots. No tenemos tiempo de hacer estudios, pero ayuda pensar en el propio riesgo para hacerlo bien."

Our founder @alixdunn.com is co-hosting this session on data centers with @justinhendrix.bsky.social CEO + Editor of @techpolicypress.bsky.social. Together we’ll hear from Paz Peña, Mozilla Fellow; @pjarandia.bsky.social, Pulitzer Fellow; and @tpang.bsky.social, Impact Editor at Lighthouse Reports.

To give a full picture, probably we have to look at how many governments around the world are finding in technology the perfect excuse to privatize and demolish our very much needed public services.

As experts point out in the piece, a universal public health system is a very complex structure full of protocols and decisions that have to be taken timely. So the causes behind this crisis are multiple and are not only related to faulty software.

Also ending or reducing the in-person meetings where doctors from different departments shared information, a critical space to decide which patients have to be prioritized.

All this was happening in parallel to privatization of services in the regional health system by Andalucía's right-wing government. Cutbacks meant that many mammographies began to be performed in mobile units by a subcontracted company.

Doctors from this hospital also describe errors by design in the software storing the mammographies, happening since it was allocated to IT giant NTT Data. Problems were flagged by radiologists working directly with the software, but they were never tackled by the officials in charge.

It is still unclear how many of these women ended up developing advanced cancer due to lack of medical care. But the organization that uncovered the scandal claims to have received more than 200 cases of affected patients whose disease was not detected in time. Some of them have already died.

Top officials from the biggest hospital in Andalucía ordered not to call by phone or send letters to the women with an inconclusive diagnosis after a mammography, since a software would do so in an automated manner. But in thousands of cases this never happened.

elpais.com/sociedad/202...

The answer to this is not simple. I personally think that we have to think bigger and be more ambitious when we investigate and explain the costs of these developments. To do so we need time, resources and to collaborate intensely between reporters with different skills.

@daniel-howden.bsky.social invited us and @joannaskao.bsky.social from the @pulitzercenter.org to think together on how journos can approach these kinds of stories. And also how we can make people care about the many implications of what probably is the biggest infra project of our modern times.

My colleague @laismartins.com puts it very clearly: "This is not a story only about infrastructure, but a story about energy, power and greed". Laís said this a few weeks ago at the session I am sharing below, at the recent iMEdD International Journalism Forum. www.youtube.com/watch?v=YKGZ...

Amassing the resources and land needed, through facilities in different countries, to power the AI tools that are making Altman & OpenAI extremely rich. I cannot think of any corporate endeavour other than Stargate that better reflects what is at stake when we talk about the AI infrastructure race.