September 2024 - nexa - server-nexa.polito.it

Social media and online video firms are conducting ‘vast surveillance’ on users, FTC finds
by Alberto Cammozzo Sept. 20, 2024

Sept. 20, 2024

<https://www.theguardian.com/technology/2024/sep/19/social-media-companies-s…> www.theguardian.com Social media and online video firms are conducting ‘vast surveillance’ on users, FTC finds Johana Bhuiyan Social media and online video companies are collecting huge troves of your personal information on and off their websites or apps and sharing it with a wide range of third-party entities, a new Federal Trade Commission (FTC) staff report on nine tech companies confirms. The FTC report published on Thursday looked at the data-gathering practices of Facebook, WhatsApp, YouTube, Discord, Reddit, Amazon, Snap, TikTok and Twitter/X between January 2019 and 31 December 2020. The majority of the companies’ business models incentivized tracking how people engaged with their platforms, collecting their personal data and using it to determine what content and ads users see on their feeds, the report states. The FTC’s findings validate years of reporting on the depth and breadth of these companies’ tracking practices and call out the tech firms for “vast surveillance of users”. The agency is recommending Congress pass federal privacy regulations based on what it has documented. In particular, the agency is urging lawmakers to recognize that the business models of many of these companies do little to incentivize effective self-regulation or protection of user data. “Recognizing this basic fact is important for enforcers and policymakers alike because any efforts to limit or regulate how these firms harvest troves of people’s personal data will conflict with their primary business incentives,” FTC chair Lina Khan said in a statement. “To craft effective rules or remedies limiting this data collection, policymakers will need to ensure that violating the law is not more lucrative than abiding by it.” The FTC is also asking that the companies mentioned in the report invest in “limiting data retention and sharing, restricting targeted advertising, and strengthening protections for teens”. Notably, the report highlights how consumers have little control over how these companies use and share their personal details. Most companies collected or inferred demographic information about users such as age, gender and language. Some collected information about household income, education and parental and marital status. But even when this type of personal information was not explicitly collected, some companies could analyze user behavior on the platform to deduce the details of their personal lives without their knowledge. For instance, some companies’ user interest categories included “baby, kids and maternity”, which would reveal parental status, or “newlyweds” and “divorce support”, which would reveal marital status. This information was then used by some companies to tailor what content people saw to increase engagement on their platforms. In some cases, that demographic information was shared with third-party entities to help target them with more relevant advertisements. Whatever product was in use, it was not easy to opt out of data collection, according to the FTC. Nearly all the companies said they fed personal information to automated systems, most often to serve content and advertisements. On the flipside, almost none of them offered “a comprehensive ability to directly control or opt-out of use of their data by all Algorithms, Data Analytics, or AI”, per the report. Several firms say it’s impossible to even compile a full list of who they share your data with. When the companies were asked to enumerate which advertisers, data brokers or other entities they shared consumer data with, none of these nine firms provided the FTC with a full inventory. The FTC also found that despite evidence that children and teens use many of these platforms, many of the tech companies reported that, because their platforms are not directed at children, they do not need different data-sharing practices for children under 13 years of age. According to the report, none of the companies reported having data-sharing practices that treated the information collected about and from 13- to 17-year-olds via their sites and apps differently than adult data, even though data about minors is more sensitive. The FTC called the companies’ data-minimization practices “woefully inadequate”, finding that some of the companies did not delete information when users requested it. “Even those Companies that actually deleted data would only delete some data, but not all,” the report stated. “That is the most basic requirement,” said Mario Trujillo, a staff attorney at the Electronic Frontier Foundation. “The fact that some weren’t doing that even in the face of state privacy laws that require it proves that stronger enforcement is needed, especially from consumers themselves.” Some of the firms have disputed the report’s findings. In a statement, Discord said the FTC report was an important step but lumped “very different models into one bucket”. “Discord’s business model is very different – we are a real-time communications platform with strong user privacy controls and no feeds for endless scrolling. At the time of the study, Discord did not run a formal digital advertising service,” Kate Sheerin, Discord’s head of public policy in the US and Canada, said in a statement. A Google spokesperson said the company had the strictest privacy policies in the industry. “We never sell people’s personal information and we don’t use sensitive information to serve ads. We prohibit ad personalization for users under 18 and we don’t personalize ads to anyone watching ‘made for kids content’ on YouTube,” said Google spokesperson, José Castañeda. The other firms either did not provide an on-the-record comment or did not immediately respond to a request for comment. However, if companies dispute the FTC’s findings, the onus is on them to provide evidence, says the Electronic Privacy Information Center (Epic), a Washington DC-based public interest research organization focused on privacy and free speech. “I used to work in privacy compliance for companies, and let’s just say I believe absolutely nothing without documentation to back up claims,” said Epic global privacy counsel, Calli Schroeder. “And I agree with the FTC’s conclusion that self-regulation is a failure. Companies have repeatedly shown that their priority is profit and they will only take consumer protection and privacy issues seriously when failing to do so affects that profit.”

2 1

Seminario Protection of AI and virtual objects in virtual spaces - 27 settembre ore 14h30 (CEST)
by Cristiana Sappa Sept. 20, 2024

Sept. 20, 2024

Carissimi/e, Mi permetto di segnalare un seminario che si svolgerà a IESEG, Campus Parigi (La Defense) il giorno 27 settembre, dalle 14h30 alle 16h00. Il tema dell'incontro è: Protection of AI and virtual objects in virtual spaces. La prospettiva sarà quella del diritto giapponese e cinese. E' possibile partecipare in presenza, dopo previa informazione spedita entro il 26 settembre all'indirizzo: c.sappa(a)ieseg.fr Oppure è possibile partecipare da remoto, tramite il link Zoom che segue: https://ieseg.zoom.us/j/83595970311 Mi tengo a disposizione per informazioni ulteriori. Un saluto, Cristiana

1 0

Why wordfreq will not be updated
by Giacomo Tesio Sept. 20, 2024

Sept. 20, 2024

The wordfreq data is a snapshot of language that could be found in various online sources up through 2021. There are several reasons why it will not be updated anymore. # Generative AI has polluted the data I don't think anyone has reliable information about post-2021 language usage by humans. The open Web (via OSCAR) was one of wordfreq's data sources. Now the Web at large is full of slop generated by large language models, written by no one to communicate nothing. Including this slop in the data skews the word frequencies. [...] As one example, [Philip Shapira reports]() that ChatGPT (OpenAI's popular brand of generative language model circa 2024) is obsessed with the word "delve" in a way that people never have been, and caused its overall frequency to increase by an order of magnitude. ## Information that used to be free became expensive [...] Even if X made its raw data feed available (which it doesn't), there would be no valuable information to be found there. Reddit also stopped providing public data archives, and now they sell their archives at a price that only OpenAI will pay. ## I don't want to be part of this scene anymore wordfreq used to be at the intersection of my interests. I was doing corpus linguistics in a way that could also benefit natural language processing tools. The field I know as "natural language processing" is hard to find these days. It's all being devoured by generative AI. Other techniques still exist but generative AI sucks up all the air in the room and gets all the money. It's rare to see NLP research that doesn't have a dependency on closed data controlled by OpenAI and Google, two companies that I already despise. wordfreq was built by collecting a whole lot of text in a lot of languages. That used to be a pretty reasonable thing to do, and not the kind of thing someone would be likely to object to. Now, the text-slurping tools are mostly used for training generative AI, and people are quite rightly on the defensive. If someone is collecting all the text from your books, articles, Web site, or public posts, it's very likely because they are creating a plagiarism machine that will claim your words as its own. So I don't want to work on anything that could be confused with generative AI, or that could benefit generative AI. https://github.com/rspeer/wordfreq/blob/master/SUNSET.md [1] https://pshapira.net/2024/03/31/delving-into-delve/

1 0

La regolamentazione UE su IA è frammentata, incerta ed incoerente
by Antonio Sept. 20, 2024

Sept. 20, 2024

"Senza regole coerenti, l'UE rischia di perdersi i due momenti fondamentali nel processo di innovazione dell'IA. Il primo riguarda lo sviluppo dei modelli "aperti", che sono resi disponibili gratuitamente per essere utilizzati da chiunque, modificati e sviluppati, moltiplicandone così i benefici e offrendo opportunità sociali ed economiche. ... Il secondo riguarda i più recenti modelli "multimodali", che operano fluidamente tra testo, immagini e audio, e che consentiranno il prossimo salto in avanti nell'IA. La differenza tra modelli di solo testo e modelli multimodali è paragonabile alla differenza tra avere un solo senso e averli tutti e cinque." La lettera aperta è qui: https://euneedsai.com/it Confesso che non ho capito a cosa si riferiscano esattamente. A.

5 5

Why the AI Hype is another Tech Bubble - Floridi
by Enrico Nardelli Sept. 20, 2024

Sept. 20, 2024

This article argues that the current hype surrounding artificial intelligence (AI) exhibits characteristics of a tech bubble, based on parallels with five previous technological bubbles: the Dot-Com Bubble, the Telecom Bubble, the Chinese Tech Bubble, the Cryptocurrency Boom, and the Tech Stock Bubble. The AI hype cycle shares with them some essential features, including the presence of potentially disruptive technology, speculation outpacing reality, the emergence of new valuation paradigms, significant retail investor participation, and a lack of adequate regulation. The article also highlights other specific similarities, such as the proliferation of AI startups, inflated valuations, and the ethical concerns associated with the technology. While acknowledging AI's transformative potential, the article calls for pragmatic caution, evidence-based planning, and critical thinking in approaching the current hype. It concludes by offering some recommendations to minimise the negative impact of the impending bubble burst, emphasising the importance of focusing on sustainable business models and real-world applications, maintaining a balanced perspective on AI's potential and limitations, and supporting the development of effective regulatory frameworks to guide the technology's design, development, and deployment. Luciano Floridi https://papers.ssrn.com/sol3/papers.cfm?abstract_id=4960826 1. Introduction The tech industry has seen several dramatic booms and busts over the past few decades. These periods, often called “tech bubbles” (see section 2), have significantly impacted economies, industries, and individual fortunes, influencing political economies and even international relations, especially between the US and China. In this article, I analyse the morphology of tech bubbles, extract some invariants and lessons, and argue that the current hype surrounding Artificial Intelligence (AI) may be considered a tech bubble leading to a new AI Winter. In section 2, I briefly analyse what a tech bubble is. In section 3, I outline some of the most notable tech bubbles, examining their causes, characteristics, and consequences. I summarise some criteria to identify a tech bubble in section 4, and argue, in section 5, that the AI hype qualifies as a tech bubble, and that its bursting may lead to a new AI Winter. In section 6, I conclude by looking into what could be done to prevent the worst consequences of such a scenario. -- -- EN https://www.hoepli.it/libro/la-rivoluzione-informatica/9788896069516.html ====================================================== Prof. Enrico Nardelli Past President di "Informatics Europe" Direttore del Laboratorio Nazionale "Informatica e Scuola" del CINI Dipartimento di Matematica - Università di Roma "Tor Vergata" Via della Ricerca Scientifica snc - 00133 Roma home page: https://www.mat.uniroma2.it/~nardelli blog: https://link-and-think.blogspot.it/ tel: +39 06 7259.4204 fax: +39 06 7259.4699 mobile: +39 335 590.2331 e-mail: nardelli(a)mat.uniroma2.it online meeting: https://blue.meet.garr.it/b/enr-y7f-t0q-ont ====================================================== --

1 0

Generative AI Has a Visual Plagiarism Problem
by Giacomo Tesio Sept. 19, 2024

Sept. 19, 2024

Experiments with Midjourney and DALL-E 3 show a copyright minefield https://spectrum.ieee.org/midjourney-copyright Di nuovo, la scoperta dell'acqua calda se si riconosce le "AI generative" come compressioni lossy dei dati sorgente: anche un jpeg compresso ha delle aree in cui gli artefatti sono nulli. Giacomo

1 0

explosive supply chain attack on Hezbollah pagers.
by 380° Sept. 18, 2024

Sept. 18, 2024

Buongiorno, siccome ne ho lette di tutti i colori in merito a "hanno stati gli hacker col MALWARE" (anche da presunti espertoni), vale la pena analizzare i fatti cum granu salis e smetterla con la SUPERSTIZIONE. https://arstechnica.com/security/2024/09/8-dead-2700-injured-after-simultan… «11 dead, thousands injured in explosive supply chain attack on Hezbollah pagers. - Supply chain attack may be to blame.» [Nate Anderson] - Updated Sep 18, 2024 2:00 am UTC --8<---------------cut here---------------start------------->8--- A massive wave of pager explosions across Lebanon and Syria beginning at 3:30 pm local time today killed at least 11 people and injured more than 2,700, according to local officials. Many of the injured appear to be Hezbollah members, although a young girl is said to be among the dead. Anonymous officials briefed on the matter are now describing it as a supply chain attack in which Israel was able to hide small amounts of explosives inside Taiwanese pagers shipped to Lebanon. The explosive was allegedly triggered by a small switch inside the pagers that would be activated upon receiving a specific code. Once that code was received, the pagers beeped for several seconds—and then detonated. [...] The shocking—and novel—attack appears to have relied on a wave of recently imported Hezbollah pagers, according to [reporting in The Wall Street Journal]. (The group had already warned its members to avoid using cell phones due to both tracking and assassination concerns.) [...] (An [early story from a local media outlet] claimed that "the pager server was compromised, leading to the installation of a script that caused an overload. This likely resulted in the overheating of the lithium battery, which then exploded.") The WSJ quoted regional security analyst Michael Horowitz as suggesting the attack was likely caused by either 1) malware triggering the batteries to overheat/explode or 2) an actual explosive charge inserted in the devices at some point in the supply chain and then detonated remotely. [...] Late in the day, US and other officials anonymously [told the New York Times] that the attack was, in fact, caused by explosives. Israel managed to insert "as little as one or two ounces" of explosive material next to the battery inside each pager, officials said. [...] Reuters has a [similar report], using different sources, but it also notes that the president of Gold Apollo emphatically claimed today that the pagers were not made by his company in Taiwan. Rather, he said, they were made under license by a company called BAC, which then put Gold Apollo's branding on them. [...] /This breaking news story has been rewritten to incorporate updates added throughout the day./ [Nate Anderson] <https://arstechnica.com/author/nate-anderson/> [reporting in The Wall Street Journal] <https://www.msn.com/en-us/news/world/hundreds-of-hezbollah-operatives-pager…> [early story from a local media outlet] <https://www.lbcgroup.tv/news/lebanon-news/796406/understanding-the-pager-an…> [told the New York Times] <https://www.nytimes.com/2024/09/17/world/middleeast/israel-hezbollah-pagers…> [similar report] <https://news.sky.com/story/explosive-pagers-used-by-hezbollah-had-been-modi…> --8<---------------cut here---------------end--------------->8--- -- 380° (Giovanni Biscuolo public alter ego) «Noi, incompetenti come siamo, non abbiamo alcun titolo per suggerire alcunché» Disinformation flourishes because many people care deeply about injustice but very few check the facts. Ask me about <https://stallmansupport.org>.

2 2

How much energy can AI use? Breaking down the toll of each ChatGPT query - The Washington Post
by Alberto Cammozzo Sept. 18, 2024

Sept. 18, 2024

<https://www.washingtonpost.com/technology/2024/09/18/energy-ai-use-electric…> A bottle of water per email: the hidden environmental costs of using AI chatbots AI bots generate a lot of heat, and keeping their computer servers running exacts a toll. [...]

1 0

Re: [nexa] explosive supply chain attack on Hezbollah pagers.
by 380° Sept. 18, 2024

Sept. 18, 2024

380° <g380(a)biscuolo.net> writes: > siccome ne ho lette di tutti i colori in merito a "hanno stati gli > hacker col MALWARE" (anche da presunti espertoni), vale la pena > analizzare i fatti cum granu salis e smetterla con la SUPERSTIZIONE. mi stavo giusto chiedendo se ci fosse qualche precedente e a rispondermi è Christine Gadsbym, Cybersecurity CISO in BlackBerry. ...poi l'articolo si trasforma in uno SPROLOQUIO sulla "software supply chain" (con soluzioni ridicole tipo gli SBOM - Software Bill Of Material), quindi mi limito a citare le parti pertinenti al supply chain attack in oggetto. https://blogs.blackberry.com/en/2024/09/pagers-explode-supply-chain-securit… «Supply Chain Lessons from Thousands of Exploding Pagers» 09.17.24, Christine Gadsby --8<---------------cut here---------------start------------->8--- But the conversation soon shifted to another possible scenario: the pagers must have been intercepted along the supply chain so someone could place explosives inside each pager with a method of remotely triggering the blasts. Attackers have utilized [[https://simple.wikipedia.org/wiki/Yahya_Ayyash][similar attacks]] before, although not to this scale. --8<---------------cut here---------------end--------------->8--- COMUNQUE: il precedente indicato è del 1996 e riguarda il sabotaggio di UN telefono cellulare nel quale furono introdotti 15 grammi di esplosivo con un sistema di detonazione remoto. --8<---------------cut here---------------start------------->8--- Kamal Hamad asked his nephew for the phone, which he later returned. It was almost certainly at this moment that the 2oz radio-controlled bomb was inserted. --8<---------------cut here---------------end--------------->8--- (https://www.independent.co.uk/news/world/how-the-phone-bomb-was-set-up-1323…) Quindi non certo un supply chain attack, ma la tecnica pare la stessa identica di circa 30 anni fa, non proprio una novità per le agenzie di "intelligence", insomma. Saluti, 380° -- 380° (Giovanni Biscuolo public alter ego) «Noi, incompetenti come siamo, non abbiamo alcun titolo per suggerire alcunché» Disinformation flourishes because many people care deeply about injustice but very few check the facts. Ask me about <https://stallmansupport.org>.

2 4

Can LLMs Generate Novel Research Ideas? A Large-Scale Human Study with 100+ NLP Researchers
by Fabio Alemagna Sept. 18, 2024

Sept. 18, 2024

«By recruiting over 100 NLP researchers to write novel ideas and blind reviews of both LLM and human ideas, we obtain the first statistically significant conclusion on current LLM capabilities for research ideation: we find LLM-generated ideas are judged as more novel (p < 0.05) than human expert ideas while being judged slightly weaker on feasibility.» https://arxiv.org/abs/2409.04109v1

1 0