Notices by Alecs Ștefănescu (catileptic@chaos.social)

i'm relieved to the point to tears to read @ali's essay:

"I’m no longer interested in encouraging the design of more human-centered versions of these murderous technologies, or to inform the more humane administration of complex algorithmic systems that separate families, bomb schools and neighborhoods, that force people out of their homes or onto the streets, or that deny medical care at the moment people need it most."

https://ali-alkhatib.com/blog/fuck-up-ai

I contributed to an investigation, for the Romanian publication Context, on how the Ministry of Exterior acquired software to profile and track journalists and influences, online. Why? Because they're trying to "combat #disinformation". With AI, of course.

Based on FOIA answers, I found out that the software is not operational one year after being bought. It also uses crude techniques, very different from what the Ministry publicly claims they're doing.

https://context.ro/dosare-personale-pentru-jurnalisti-si-influenceri-ministerul-de-externe-a-cheltuit-10-milioane-de-lei-pe-un-soft-cu-ai-care-sa-urmareasca-dezinformare-si-emotii-nefunctional-la-un-an-de-la-l/

I'm looking for texts (ideal: essays, but will gladly take books as well) that discuss what hierarchy is. I'm not really looking for deep-dives into how hierarchy has emerged. I'm just curious about attempts to define what it is and what it isn't.

I've already read David Graeber's chapter that explains hierarchy through two anthropological concepts, of joking relations & relations of avoidance: https://davidgraeber.org/articles/manners-deference-and-private-property-or-elements-for-a-general-theory-of-hierarchy/

Any other recommendations?

#philosophy #anthropology

98.06% of the votes in #Moldova have been counted and Maia Sandu, the current, pro-EU president, has been re-elected with 54.47% of votes (for now).

A lot of news around the DDoS aimed at @internetarchive mentions the group that claimed the attack, SN_BLACKMETA, as being "pro-Palestinian".

I wanted to check what the evidence behind this claim is. The most comprehensive article I could find was by Radware https://www.radware.com/security/threat-advisories-and-attack-reports/six-day-web-ddos-attack-campaign/ . The relevant part in the section titled "The Rise and Unfolding of SN_BLACKMETA".

Spoiler: Radware is headquartered in Tel Aviv. But anyway, let's see the evidence.

1/3

A group of almost 30 digital rights and LGBTQIA+ NGOs addressed an open letter to the CEOs of Meta, X, Reddit, Snap and YouTube, urging them to take proactive and firm measures to curb the damage of non-consensual sexually explicit material generated with AI.

This kind of material has been shown to disproportionately use the identities and likeness of women, LGBTQIA+ folx and racialised individuals. Their open letter links to studies that map the phenomenon.

https://weareultraviolet.org/wp-content/uploads/2024/09/SIGN-ON-LETTER-DRAFT_-AI-Demands-to-Fight-Gender-Bias-Sexism-and-Misogyny-3.pdf

@aral @JulianOliver

this is a great resource! :) (also relevant to me, personally, since i host my own websites). Thank you for linking it!

If you have the broad strokes of what's missing in your mind or in your notes, Aral, I would be happy to volunteer some time and writing, to get them on paper and update this piece of writing. :) Let me know if you think it could work.

Border authorities in China can search electronic devices as well as collect emails, pictures, and videos of foreigners entering the country. No warrant needed. These searches can be conducted in case of emergencies - although it seems like there are no checks & balances in place to reign in or even track when this is being done. Or why.

https://english.kyodonews.net/news/2024/07/1e30a810e2ba-china-starts-smartphone-inspections-to-boost-anti-espionage-efforts.html

what do you use when you want to plot values in a chart quickly and neatly?

i don't mean "for production" or "for publishing". i just mean: you have a data set and you want to see lines on a line chart for yourself, just to get an intuition about what the data shows.

curious about the go-to solution of other folx when it comes to quick #dataviz.

ahead of the vote on the new proposal for #ChatControl, worth remembering that the mass-scanning of private communications, invalidating end-to-end encryption, weakening the security of our devices via client-side scanning and treating EU citizens as potentially culpable by default are *not* viable solutions to *any* issue.

@edri published a more detail rebuttal: https://edri.org/our-work/be-scanned-or-get-banned/

wrote about the excellent & extremely quotable ChatGPT is Bullshit paper on my blog: https://catileptic.tech/posts/chatgpt-is-bullshit/

kind reminder that #llm technology is not truth-seeking, that it merely reproduces without attempting to understand, that the made-up content is an intentional effect of how the algorithm works, that #ai kills and that #chatgpt puts large amounts of natural resources to waste only to feed us bullshit

Pro-Palestinian protests in Romania have been met with hostility and control from the police forces.

From banning banners with slogans that do not infringe on rights, and neither do they incite to violence, to harassing protesters, to summoning people who intended to protest to the police station for "friendly advice", we see Ro police taking sides. The side of genocide.

ActiveWatch also condemns the silence of the press, which only covered these abuses sparingly: https://activewatch.ro/articole/fantomele-militiei-si-securitatii-bantuie-politia-si-jandarmeria/

The Ministry doesn't have any competence to decide what propaganda & disinformation is. The end result may be that legitimate speech gets classified as malicious, or that truly harmful speech gets classified as legit. Both are dangerous.

There *could be* a collaboration (and knowledge sharing) between Gov. & civic society & academia, under the #DSA. But the Ro #DSC is adamant in not reaching out and involving these sectors. We also don't know which ones the competent authorities are. 3/4

Last but not least, there is no transparency, nor feedback.

A user doesn't find out whether the content they reported was forwarded to the adequate VLOP. We don't even get to know how many reports the Ministry processed, who they forwarded the reports to and whether action was taken to remove the content.

The form asks citizens to provide an e-mail address, but doesn't mention why, doesn't indicate what it will be used for (hello, #GDPR).

It's all just dust in the wind (of elections) 4/4

The Romanian Ministry of Digitalization released a form that people can use to report malicious #deepfake materials. What counts as malicious? Fraud, propaganda, disinformation.

We went into detail about how misguided this is, in a Romanian article over on the @apti website: https://apti.ro/ordinul-mcid-deepfake-praf-cu-iz-electoral .

But, because this matter concerns me deeply, here's a break-down in English (especially relevant if you're following the #DSA implementation): 1/4

This ignores the main problem.

98% of #deepfake material is pornographic. 99% of the victims of deepfake pornography identify as women.

In Ro, journalists & activists have been adamant about showing the violence against women perpetrated online and aided by technology.

Yet, the Gov. chose to focus on 2% of all deepfakes, thus creating a false narrative that this tech is used mainly to influence elections and commit fraud - when, in fact, the majority of victims are being ignored. 2/4

OpenAI has partnered with StackOverflow to use the content on the site to train ChatGPT. Some users who have posted highly-rated answers have tried to either delete or edit them. But they noticed that the moderators changed their edits back and their old answers re-appeared. Then the moderators banned them for 7 days.

Oh and while we're on the subject, Valve also removed the ban on AI-generated art on Steam.

Anything we put up in these walled gardens will be monetized.

https://www.tomshardware.com/tech-industry/artificial-intelligence/stack-overflow-bans-users-en-masse-for-rebelling-against-openai-partnership-users-banned-for-deleting-answers-to-prevent-them-being-used-to-train-chatgpt

Brett Scott published a really good piece on the cost of "cashlessness".

"Money is political, and that’s the starting point of the economics of cash."

https://www.asomo.co/p/the-cost-of-cashlessness

It's especially important since the only *other* time I hear people defending cash is when they fear a take-over from a shady, world-wide organization. And it's not that they want freedom, but they want another shady, world-wide organization to *not* lose its foothold.

I'm making a list of Mastodon and other Fediverse instances either hosted in Romania, registered on a .ro domain or with content that is mostly Romanian (regardless of where they are hosted).

The reason behind this: the Ro digital service coordinator added some extra hoops on top of what DSA requires. We're trying to figure out who will be affected.

If anyone knows such instances, please leave them in a public or private toot (or boost / tag folx who can help) <3

I already know of mstdn.ro

Amnesty (who I hope will join the Fediverse soon) published an interview where I repeat, as a little broken record, that all tech is political. There's no such thing as neutral tech.

“My approach to technology is informed by political struggles.”

It features spyware, ChatControl, and the Amnesty Digital Forensics Fellowship.

https://www.amnesty.org/en/latest/campaigns/2024/01/my-approach-to-technology-is-informed-by-political-struggles/