| David Garcia via nettime-l on Fri, 20 Nov 2026 18:42:08 +0100 (CET) |
[Date Prev] [Date Next] [Thread Prev] [Thread Next] [Date Index] [Thread Index]
| Re: <nettime> Falling into the Breach |
Another item on your list on the road to Hell I would add is:
- AI blurs and so weakens lines of accountability.
In his article on documentary NAZA (about Israel's AI enabled killing machine and system for calculating "collateral" damage) Eyal Weizman, writing in the London Review of Books, describes how Daniel Reisner, chief legal adviser in the Israeli army before he died in July, became frustrated with the vagueness of the concept of proportionality in warfare. Proportionality, he argued, doesn’t ‘tell us what to include in the calculation, what is the equation and what is the exchange rate’?
He thought armies needed simple formulas and exact numbers to be able to act quickly and efficiently ‘in real-time situations’. He was trying to develop what he described to Weizman 15 years ago, as ‘mathematical tools to tackle the problem. He believed that letting computers make the calculations was more *ethical* than leaving it to humans, because machines aren’t subject to human bias and aren’t corrupted by the desire for revenge, cruelty or excess.
One of the interviewee/operatives in the movie shares Reisner's position, telling co-director Abraham that ‘it’s better that the AI takes the decision rather than a person whose friend was just killed.’
David Garcia On 2026-11-18 16:03, Felix Stalder via nettime-l wrote:
I think AI plays into this on multiple levels, but I'm not sure the attack level is the most significant at the moment.- Bosses have been convinced that they could fire IT staff, replacing them with AI, leaving many orgs critically understaffed and with fading institutional memory.- Weakening of regulatory oversight, and a culture of no-accountability, even before Trump introduced rampant cronyism. No software company was ever fined for a security breach or system failure (as far as I know).- AI has been able to uncover many long existing vulnerabilities. Security by obscurity no longer works, putting a lot of pressure on orgs to patch their systems. Probably their IT departments warned about some of those issues for a long time, but now AI says it, so management listens.- AI shortens the time between vulnerability detected and exploit deployed, putting even more pressure in the orgs.- Patches are vibe-coded and/or rolled out at high speeds with insufficient testing, making the systems unstable in unforeseen ways.For well-maintained clearly bounded projects such as the Linux Kernel, this is a manageable process that might result in improved software, for large, sprawling infrastructures, where many systems are hobbled together, this is a real problem.all the best. Felix On 18/11/2026 19:34, Brian Holmes via nettime-l wrote:Ted wrote:"Institutionalist analyses read well because they can seem on their face soconcrete and specific. But in this one I’m definitely taking theculturalist view. It doesn’t negate the things you’ve mentioned (and legal frameworks are definitely institutionalist in nature), it just situates them more meaningfully in something like a Geertzian thick description."I am definitely in favor of the thick description. I am currently reading about the concept of pre-compliance (aka anticipatory obedience) which is used to analyze the combined phenomena of information control in the press and direct repression by civil society actors (museum boards, university officials, Silicon Valley firms etc). Pre-compliance has a long history, stretching back through 9/11 to the Cold War. So yeah, no objection at all,we're just onto different aspects.Probably I am more ignorant than many of you as to the geeky stuff. But since we have before our eyes what looks like a big pre-compliance event in the press, I am fascinated by what is not being talked about: the technicaldetails of AI-enabled hacking, plus the more familiar structure of arms-length support by states for those kinds of cyber-attacks.Interestingly the NY Times has only one mention of ARTEX (a month ago), andthey seem to have said nothing about the Optimizely intrusion, anothermajor hack that took place the day before yesterday. While the state failsand the media obfuscates, the situation grows more extreme.The development of AI has created an emergency situation: not rogue agents but super-empowered human actors. The Trump administration is choking off news about it and the media are complying with state control. Maybe they'reafraid of popular panic? On the ground and even in the air, we are all sitting ducks for whatever may happen. curiously, BrianOn Wed, Nov 18, 2026 at 10:20 AM GM - tedbyfield <tedbyfield@gmail.com>wrote:I’m not sure how you could read my litany of breakdowns and could go on to say so affirmatively that it misses a few key things, but you do you. The legal frameworks you mention are important, sure, but to say they’re key suggests that the rule of of law prevails — but *the defining feature* of Trump 2 is that the rule of law does *not* apply. Innocent parties are harassed with lawfare, guilty parties are rewarded and honored, existing laws are ignored, and imaginary laws are imposed ruthlessly — all accordingto a ‘logic’ so wildly erratic that to call it hormonal would be toogenerous. If a company actively violated one of the legal frameworks you mention but did so in ways that actively supported the whims of Trump & Co, the biggest risk they’d run by far would be the inviolable principle that everything Trump touched dies sooner or later. So, sure, those laws are a factor in companies’ decision to remain silent, but they won’t be the first thing that comes to mind in the C-suite (or even the counsel’s office, I’dwager). And I say that on the basis of decades of paying pretty closeattention to how various entities in digital security have navigated theshifting boundaries of compliance, disclosure, etc.Also: the litany of breakdowns in my last mail left out a *huge* factor in this context — the Trumpian attack on higher ed. The resulting pressure in HE to keep your head down and your moiuth shut has become excruciating across the board, but in sensitive areas — constitutional law, for example — it doesn’t matter which way someone throws the dice, they *will* lose. If they don’t run afoul of Trump & Co, they’ll get spanked by admins who live in terror of the Eye of Sauron will fall on their institution, orincur the mistrust of their colleagues and peers.Historically, compsec has relied *very* heavily on academic research for a variety of reasons — clearly structured areas of expertise, an equitable issue-driven approach, rigorous standards, noncommercial incentives, legal protections related to educational contexts, and so on. That old-school (so to speak) system — exemplified by orgs like CERT at CMU, Spafford’s COAST at Purdue, UC Davis’s Computer Security Lab, etc — has been breaking down for quite a while, but now it seems to be dead. In fields like this, HE researchers were like honeybees who could cross boundaries of every kind, building networks and encouraging productively open exchange, but now not.More generally, I’ve been noticing more and more that discussions likethis tend to break down on institutionalist vs culturalist lines.Institutionalist analyses read well because they can seem on their face soconcrete and specific. But in this one I’m definitely taking theculturalist view. It doesn’t negate the things you’ve mentioned (and legal frameworks are definitely institutionalist in nature), it just situates them more meaningfully in something like a Geertzian thick description.Ted On 18 Nov 2026, at 10:02, Brian Holmes via nettime-l wrote:Folks, I totally agree that the infrastructure in the US is old and the administration is vastly incompetent, making even very important systems like air traffic control uniquely vulnerable. But the observation missestwo key things about the present situation.The first is that there is very likely a reason for the current confusion: it's because US officials and their private contractors are legally barred from speaking about cybersecurity events, due to legal frameworks likethese: <. . .>-- | |||||||||||||||| http://felix.openflows.com | | |||||||||| https://tldr.nettime.org/@festal | | for secure communication, please use signal |
-- # distributed via <nettime>: no commercial use without permission # <nettime> is a moderated mailing list for net criticism, # collaborative text filtering and cultural politics of the nets # more info: https://www.nettime.org # contact: nettime-l-owner@lists.nettime.org