While technology itself is apolitical the direction of technological development and control of technology is very much a political topic. For example nobody can deny that the decision to invest in research of bioweapons instead of vaccines is a political decision.
In the context of the Western world today this decision is left to the whims of the ultra wealthy who manage fortunes bigger than some countries’ budgets. This leads to no democratic control and society absorbing all externalities while the ultra rich reap the benefits.

This is not a recent problem – it dates back to the days of the industrial revolution, giving raise to legendary figures like Ned Ludd and the movement Luddism. Movement that is now going trough a renaissance exactly because their ideas are so relevant. And while the technology has changed drastically and has in turn changed the world, the problems behind that lack of democratic control remain mostly unaddressed.
The technology on focus today is AI, how it impacts the environment, our lives and how it is used to take lives.

There is a plan to build a new AI datacenter near Plovdiv. The plan is proposed by the German company Brinell Compute GmbH and while it may seem like a great opportunity to build our local computational capacity, there are some major questions that no one is asking.

How big will the datacenter be? What will that mean for the electricity grid? And who will pay for a potential capacity build out?
These questions has been asked before and for a good reason. A lot of households in America find their electricity bills jumping because they are forced by their electricity providers to absorb the cost of new compute capacity intended to be used for datacenters.

What is the environmental impact?
Datacenters are famous for consuming water and producing copious amounts of noise pollution and heat. How is that going to be handled in a region that is already water deficient and lethally hot during the summer?

What economic impact will the datacenter have on the surroundings?
A lesser-known fact is that once a datacenter is built it requires only minimal personnel to operate. Meaning this construction will most likely not lead to great new job opportunities, and the revenue created will be siphoned back to the company that owns the datacenter.

This doesn’t seem like a build out of local capacity, but rather like export of pollution and plunder of natural resources. And worst comes yet.

The truly horrifying part of AI is its use in war. More precisely AI used for “target generation”.
In war target generation is the process of selecting objects to attack, suck as military bases, military personnel, etc.
What is so horrifying about this can be seen in the ongoing genocide in Gaza and a system called “Lavender”. Developed to solve the “human bottleneck for both locating the new targets and decision-making to approve the targets” Lavender has been used to great extent and great “success” by the Israeli government in accord with “Where’s Daddy?” – another automated system intended to track targets and carry out bombings when the target is in their place of residence.
Success of course is in quotes here because in this case it means the indiscriminate killing of more than 75000 people. Specifically the fact that “70% of the Palestinians killed in residential buildings or similar housing were women and children” cannot be ignored when we know about systems like “Where’s Daddy?”.

But what about hallucinations? We all know AI is prone to hallucinating, making up facts, confidently blundering a calculation or failing to count the number of “r” in a “strawberry”.
This should be very worrying when planning for example a covert military operation, but that is only dependent on the goal, if the goal on the other hand is to bulldoze a whole city while killing, maiming and displacing its population, hallucinating becomes only one more way to generate more targets. And targets it generates, with error policies such as “for every junior Hamas operative that Lavender marked, it was permissible to kill up to 15 or 20 civilians” and “There was no ‘zero-error’ policy. Mistakes were treated statistically“. There must be not a drop of suspicion that the system’s hallucinations were just used to further the goal of genocide.

This amalgamation of government surveillance, technological capabilities and most importantly genocidal intent is exactly what enables the use of technologies like Lavender to their full, ugly, destructive, homicidal potential.