- nexa - server-nexa.polito.it

On the Opportunities and Risks of Foundation Models
by Giuseppe Attardi Aug. 23, 2021

Aug. 23, 2021

Rapporto di 200 pagine dal nuovo Center for Research on Foundation Models di Stanford. I Foundation Model sono una generalizzazione dei Language Models (es. GPT-3), allenati su una larga scala di dati, adattabili a un’ampia varietà di compiti. In particolare il rapporto affronta gli impatti sociali di tai modelli, ossia: (i) the harms with respect to inequity and misuse, (ii) the impact with respect to the economy and environment, and (iii) the broader considerations with respect to the law and ethics https://arxiv.org/pdf/2108.07258.pdf <https://arxiv.org/pdf/2108.07258.pdf> — Beppe

2 1

Conspiracy theories aside, there is something fishy about the Great Reset | openDemocracy
by J.C. DE MARTIN Aug. 23, 2021

Aug. 23, 2021

È piuttosto interessante vedere come il modello “multistakeholder”, nato - credo - in ambito “Internet governance”, stia diventando il paradigma per la “governance” del mondo tout court: https://www.opendemocracy.net/en/oureconomy/conspiracy-theories-aside-there… jc (Sent from my wireless device; please excuse brevity and typos (if any))

1 0

In a world first, South Africa grants patent to an artificial intelligence system
by don Luca Peyron Aug. 23, 2021

Aug. 23, 2021

The patent—which was filed by an international team of lawyers and researchers led by the University of Surrey's, Professor of Law and Health Sciences, Ryan Abbott —is for a food container based on fractal geometry. This container was designed and created by an AI called DABUS ("device for the autonomous bootstrapping of unified sentience"). https://www.surrey.ac.uk/news/worlds-first-patent-awarded-invention-made-ai… https://theconversation.com/in-a-world-first-south-africa-grants-patent-to-… _________________________ don Luca Peyron Direttore Ufficio per la Pastorale Universitaria Arcidiocesi di Torino www.universitari.to.it via XX settembre 83, Torino tel. 011 5156239

1 0

Apple photo-scanning plan faces global backlash from 90 rights groups | Ars Technica
by J.C. DE MARTIN Aug. 22, 2021

Aug. 22, 2021

https://arstechnica.com/tech-policy/2021/08/apple-photo-scanning-plan-faces… (Sent from my wireless device; please excuse brevity and typos (if any))

1 0

Surveillance and capture: Two models of privacy (Philip E. Agre)
by 380° Aug. 21, 2021

Aug. 21, 2021

Buongiorno, grazie infinite a J.C. De Martin che con la sua segnalazione [1] della /mitica/ quanto negletta persona di Philip E. Agre mi ha dato l'opportunità di leggere il suo *splendido* saggio «Surveillance and capture: Two models of privacy», 1994 disponibile sia su Sci-Hub: https://sci-hub.st/10.1080/01972243.1994.9960162 che qui, ripubblicato in The New Media Reader nel 2003: https://blogs.commons.georgetown.edu/cctp-506-spring2018/files/2015/12/51-a… Venti anni prima che Vincent Mosco descrivesse la vendita delle informazioni ai pubblicitari come "surveillance capitalism" e 25 anni prima di Shoshana Zuboff col suo famosissimo libro, Agre descrive le questione della privacy come dialettica tra metafore visuali (surveillance) metafore linguistiche (capture): --8<---------------cut here---------------start------------->8--- Abstract: Two models of privacy issues are contrasted. The surveillance model employs visual metaphors (e.g., "Big Brother is watching") and derives from historical experiences of secret police surveillance. The less familiar capture model employs linguistic metaphors and has deep roots in the practices of applied computing through which human activities are systematically reorganized to allow computers to track them in real time. The capture model is discussed with reference to systems in numerous domains. --8<---------------cut here---------------end--------------->8--- Ecco i due modelli "affiancati" (estratto da pag. 23): --8<---------------cut here---------------start------------->8--- Let us review the definitions of these models, recalling once again that they are intended as metaphor systems and not as mutually exclusive categories: (1) The surveillance model employs visual metaphors, most famously Orwell's "Big Brother is watching you"; the capture model employs linguistic metaphors by means of various grammars of action. (2) The surveillance model emphasizes nondisruptive, surreptitious data collection; the capture model describes the readily apparent instrumentation that entails the reorganization of existing activities. (3) The surveillance model is concerned to mark off a "private" region by means of territorial metaphors of "invasion" and the like; the capture model portrays captured activities as being constructed in real time from a set of institutionally standardized parts specified by the captured ontology. (4) The surveillance model depicts the monitoring of activity as centrally organized and presumes that the resulting information is centrally stored; the capture model emphasizes the locally organized nature of contests over the capture process and their structuring within particular institutional contexts. (5) The surveillance model takes as its prototype the malevolent political activities of state organizations; the capture model takes as its prototype the quasi-philosophical project of ontological reconstruction undertaken by computer professionals in private organizations. --8<---------------cut here---------------end--------------->8--- É un saggio /molto/ denso, con numerosi richiami a ricerche e progetti informatici e anche a questioni filosofiche con riferimenti diretti ad Heidegger e - indirettamente - a Foucault, quindi un saggio praticamente *impossibile* da riassumere perché non-riassumibile. Tuttavia, tra le molte cose trattate, mi permetto di evidenziare il capitolo "Grammars of Action" (pag. 8) - che nel modello "capture" sono centrali - ed in particolare di quel capitolo: --8<---------------cut here---------------start------------->8--- [...] The capture model describes the situation that results when grammars of action are imposed upon human activities, and when the newly reorganized activities are represented by computers in real time. It is convenient to subdivide this process into a five-stage cycle. This division is, of course, a great oversimplification: The phases frequently operate concurrently, advances in one phase may force revision of the work done in an earlier phase, and work in each stage draws on a wide range of sociotechnical advances not necessarily related to the other stages. The five main stages are as follows: • Analysis. Somebody studies an existing form of activity and identifies its fundamental units in terms of some ontology (entities, relations, functions, primi tive actions, and so forth). This ontology might draw on the participants' terms for things, or it might not. Programming languages and systems analysis methodologies frequently supply basic ontologies (objects, variables, relations) upon which domain-specific ontologies can be built. The resulting ontologies are sometimes standardized across whole institutions, industries, or markets. • Articulation. Somebody articulates a grammar of the ways in which those units can be strung together to form actual sensible stretches of activity. This process can be complicated, and it often requires revision of the preceding ontological analysis. It is typically guided by an almost aesthetic criterion of obtaining a complete, closed, formally specified picture of the activity. • Imposition. The resulting grammar is then given a normative force. The people who engage in the articulated activity are somehow induced to organize their actions so that they are readily "parsable" in terms of the grammar. The "somehow" is typically both social (explicit procedures backed up by certain relations of authority) and technical (whether through machinery or simply through physical barriers); participants in the activity may or may not participate in the process and may or may not resist it. Institutions frequently impose grammars on activities for reasons other than real-time capture—for example, for security, efficiency, protection from liability, and simple control. • Instrumentation. Social and technical means are provided, whether through paperwork or machinery, and potentially with a complex division of labor, for maintaining a running parse of the ongoing activity. This phase may coincide with the imposition phase, or it may follow by years or decades. Afterward, the participants begin, of necessity, to orient their activities toward the capture machinery and its institutional consequences. • Elaboration. The captured activity records, which are in economic terms among the products of the reorganized activity, can now be stored, inspected, audited, merged with other records, subjected to statistical analysis, employed as the basis of Pareto optimization, and so forth. Likewise, concurrent computational processes can use captured records to "listen to" the ongoing activities for purposes of error detection, advice giving, performance measurement, quality control, and so forth. These additional processes might arise simultaneously with the instrumentation phase, or they may accumulate long afterward. --8<---------------cut here---------------end--------------->8--- Ho la netta sensazione che questa analisi aiuti anche a comprendere /come/ l'informatica sarà uno strumento per ristrutturare i rapporti di potere tecnici, sociali, economici e politici così come li conosciamo oggi. OK smetto qui che è meglio :-D Saluti, 380° [1] https://server-nexa.polito.it/pipermail/nexa/2021-August/022080.html -- 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>.

1 1

Computer quantistici, trovato un modello che supera Google
by don Luca Peyron Aug. 21, 2021

Aug. 21, 2021

Due notizie in una: da un parte il bel passo in avanti per la comunità scientifica e non solo rispetto alla tecnica in sè, dall'altra il fatto che il comunicato stampa debba riportare il raffronto con Google. Per avere spazio in pagina? Per dire siamo Davide contro Golia? Per segnare un territorio? Per... mettete voi il seguito che intuite. Un segnale, tra i molti, per dire dove stiamo. Buona lettura dl https://www.cnr.it/it/comunicato-stampa/10520/computer-quantistici-trovato-… _________________________ don Luca Peyron Direttore Ufficio per la Pastorale Universitaria Arcidiocesi di Torino www.universitari.to.it via XX settembre 83, Torino tel. 011 5156239

1 0

Google says geofence warrants make up one-quarter of all US demands – TechCrunch
by J.C. DE MARTIN Aug. 20, 2021

Aug. 20, 2021

https://techcrunch.com/2021/08/19/google-geofence-warrants/ (Sent from my wireless device; please excuse brevity and typos (if any))

1 0

The Risks of Election Believability (or Lack Thereof) - By Rebecca T. Mercuri, Peter G. Neumann CACM, June 2021, 64(6)
by Diego.Latella Aug. 20, 2021

Aug. 20, 2021

E' stato pubblicato a giugno, ma non mi pare sia passato nella lista: The Risks of Election Believability (or Lack Thereof) By Rebecca T. Mercuri, Peter G. Neumann CACM, June 2021, 64(6) https://cacm.acm.org/magazines/2021/6/252836-the-risks-of-election-believab… Saluti Diego -- Dott. Diego Latella - Senior Researcher CNR/ISTI, Via Moruzzi 1, 56124 Pisa, Italy (http:www.isti.cnr.it [1]) FM&&T Lab. (http://fmt.isti.cnr.it) CNR/GI-STS (http://gists.pi.cnr.it) https://www.isti.cnr.it/People/D.Latella - ph: +390506212982, mob: +39 348 8283101, fax: +390506212040 =================== I don't quite know whether it is especially computer science or its subdiscipline Artificial Intelligence that has such an enormous affection for euphemism. We speak so spectacularly and so readily of computer systems that understand, that see, decide, make judgments, and so on, without ourselves recognizing our own superficiality and immeasurable naivete with respect to these concepts. And, in the process of so speaking, we anesthetise our ability to evaluate the quality of our work and, what is more important, to identify and become conscious of its end use. […] One can't escape this state without asking, again and again: "What do I actually do? What is the final application and use of the products of my work?" and ultimately, "am I content or ashamed to have contributed to this use?" -- Prof. Joseph Weizenbaum ["Not without us", ACM SIGCAS 16(2-3) 2--7 - Aug. 1986] Links: ------ [1] http://www.isti.cnr.it

1 0

Tassare la pubblicità per regolamentare le Big Tech
by Giuseppe Attardi Aug. 19, 2021

Aug. 19, 2021

Da tempo sostengo che occorre intervenire sulla pubblicità per regolamentare le Big Tech. Ora lo dice anche il premio Nobel in Economia Paul Romer. https://www.technologyreview.com/2021/06/14/1026207/taxing-digital-advertis… — Beppe

8 22

The Taliban Have Seized U.S. Military Biometrics Devices
by Alberto Cammozzo Aug. 19, 2021

Aug. 19, 2021

<https://theintercept.com/2021/08/17/afghanistan-taliban-military-biometrics/> _The Taliban have_ seized U.S. military biometrics devices that could aid in the identification of Afghans who assisted coalition forces, current and former military officials have told The Intercept. The devices, known as HIIDE, for Handheld Interagency Identity Detection Equipment, were seized last week during the Taliban’s offensive, according to a Joint Special Operations Command official and three former U.S. military personnel, all of whom worried that sensitive data they contain could be used by the Taliban. HIIDE devices contain identifying biometric data such as iris scans and fingerprints, as well as biographical information, and are used to access large centralized databases. It’s unclear how much of the U.S. military’s biometric database on the Afghan population has been compromised. While billed by the U.S. military as a means of tracking terrorists <https://www.army.mil/article/32609/bats_helps_id_insurgents_hostages> and other insurgents, biometric data on Afghans who assisted the U.S. was also widely collected and used in identification cards, sources said. “We processed thousands of locals a day, had to ID, sweep for suicide vests, weapons, intel gathering, etc.” a U.S. military contractor explained. “[HIIDE] was used as a biometric ID tool to help ID locals working for the coalition.” A spokesperson for the Defense Intelligence Agency referred questions to the Office of the Secretary of Defense, which did not respond to a request for comment. An Army Special Operations veteran said it’s possible that the Taliban may need additional tools to process the HIIDE data but expressed concerns that Pakistan would assist with this. “The Taliban doesn’t have the gear to use the data but the ISI do,” the former Special Operations official said, referring to Pakistan’s spy agency, Inter-Services Intelligence. The ISI has been known to work closely with the Taliban. The U.S. military has long used HIIDE devices in the global war on terror and used biometrics to help identify <https://www.scientificamerican.com/article/how-biometrics-helped-to-identif…> Osama bin Laden during the 2011 raid on his Pakistani hideout. According to investigative reporter Annie Jacobsen, the Pentagon had a goal <https://www.npr.org/2021/01/14/956705029/first-platoon-examines-how-war-on-…> to gather biometric data on 80 percent of the Afghan population to locate terrorists and criminals. <https://www.army.mil/article/47549/new_equipment_helps_troops_locate_insurg…> “I don’t think anyone ever thought about data privacy or what to do in the event the [HIIDE] system fell into the wrong hands,” said Welton Chang, chief technology officer for Human Rights First, himself a former Army intelligence officer. “Moving forward, the U.S. military and diplomatic apparatus should think carefully about whether to deploy these systems again in situations as tenuous as Afghanistan.” The Defense Department has also sought to share the biometrics data collected by HIIDE with other government agencies such as the Federal Bureau of Investigation and Department of Homeland Security. In 2011, the Government Accountability Office criticized <https://www.gao.gov/assets/a317375.html> the Pentagon for not doing enough to ensure these other surveillance agencies had easy access to the information, warning that the military “limits its federal partners’ ability to identify potential criminals or terrorists.” But the U.S. didn’t only collect information about criminals and terrorists; the government appears to also have been collecting biometrics from Afghans assisting diplomatic efforts, in addition to those working with the military. For example, a recent job posting by a State Department contractor sought to recruit a biometric technician with experience using HIIDE and other similar equipment to help vet personnel and enroll local Afghans seeking employment at U.S. embassies and consulates. The federal government has collected biometric data from Afghans despite knowing the risks entailed by maintaining large databases of personal information, especially given recent cyberattacks on government agencies and private companies. These efforts are continuing to expand. For example, a February 2020 article <https://peoiews.army.mil/aircraft-survivability-2-2/> published by the Army indicated that the service was modernizing its 20-year-old biometric processing technology and had saved more than 1 million entries in the Pentagon’s Automated Biometrics Identification System, or ABIS, which hosts HIIDE and data collected by other devices as well. “This updated database will make it more efficient for warfighters to collect, identify and neutralize the enemy,” wrote Col. Senodja Sundiata-Walker, project manager for the Pentagon’s biometrics program. President Joe Biden’s proposed budget for the Army in fiscal year 2022 seeks more than $11 million to purchase 95 new biometric collection devices expanding upon those used in Afghanistan and Iraq.

1 0