- nexa - server-nexa.polito.it

Assisi IA Act
by Norberto Patrignani Jan. 2, 2025

Jan. 2, 2025

carissim*, segnalo (in allegato per comodita') il testo completo dell'AI Assisi Act dal sito dell'Ordine dei Giornalisti: https://www.odg.it/comunicazione-e-ia-per-un-approccio-corretto-e-responsab… buon 2025! Norberto

2 1

16° Conferenza Annuale del Centro Nexa "Guerra, pace e algoritmi" | Video dell'incontro
by Nexa - Media Jan. 2, 2025

Jan. 2, 2025

Gentilissime, gentilissimi, Vi segnaliamo che è ora disponibile sul nostro canale YouTube (https://www.youtube.com/@NexaCenter) la registrazione della 16° Conferenza Annuale del Centro Nexa su Internet & Società "Guerra, pace e algoritmi". Un saluto cordiale, -- Valeria Bergantino Communication Manager Nexa Center for Internet & Society Politecnico di Torino - DAUIN Via Pier Carlo Boggio, 65/A - 10138 Torino web: https://nexa.polito.it/ mail: valeria.bergantino(a)polito.it<mailto:valeria.bergantino@polito.it> tel: 3473443585

1 0

Solove, "Privacy in Authoritarian Times"
by J.C. DE MARTIN Dec. 30, 2024

Dec. 30, 2024

Il contesto di cui parla Daniel Solove nel suono editoriale per il Boston Globe sono gli USA, per cui vari suoi ragionamenti vanno adattati al contesto EU. jc *Privacy in Authoritarian Times* Daniel Solove /Crackdowns on immigrants. Surveillance of abortion providers and abortion seekers. Harassment of critics. These maneuvers in the demagogue’s playbook would be harder to pull off with better digital privacy./ As the United States and much of the world turn back to the darkness of authoritarianism that blighted the previous century, we must remember that privacy is one of the bulwarks against the power of authoritarian governments. Unfortunately, we’re living in a more intensive surveillance society than ever before, and our privacy laws are ill equipped to protect us against government access to our personal data. In the years to come, the federal government and many state governments might engage in surveillance and data gathering as they round up immigrants, punish people for seeking, providing, or assisting abortions, and attack gender-affirming health care. The government might use personal data in its effort to retaliate against those who stand in its way. Such efforts might be assisted by mobs of vigilantes who will use personal data to dox, threaten, embarrass, and harm anyone they don’t like — much like the way many people eagerly assisted totalitarian regimes in finding “undesirables” and rooting out and punishing dissenters. Our best hope for protection is that legislators in Massachusetts and other states who are concerned about these risks take steps now to upgrade their privacy laws. Writing more than two decades ago, I critiqued how digital dossiers existed for all of us <http://ssrn.com/abstract=313301> — extensive repositories of personal data gathered by companies — and how the government could readily gain access to this data. The Fourth Amendment provides minimal to no protection for much of our personal data in the digital age. This is because of the Third-Party Doctrine, a misguided rule manufactured by the US Supreme Court in the 1970s that holds that whenever personal data is in the hands of third parties, people no longer have a reasonable expectation of privacy in it — and as a result, there’s no Fourth Amendment protection. The Third-Party Doctrine is especially problematic and ill-suited to our modern times, when so much of our personal data is in the hands of third parties such as credit card companies, search companies, cloud service providers, online merchants, data brokers, social media companies, and app creators. When the government demands that companies turn over personal data and let authorities use surveillance technologies, will companies refuse to cooperate? If history repeats itself, the answer is no. Last century, many companies supplied the Nazis with the technology <https://time.com/5290043/nazi-history-eu-data-privacy-gdpr/> and information they wanted <https://www.magellantv.com/articles/nazi-data-science-the-dark-history-that…>. Here at home, when the federal government demanded data from companies after 9/11, they readily turned it over — even when it was against their own privacy policies. Even companies that are not eager to cooperate often buckle under government pressure. [...] continua qui: https://teachprivacy.com/privacy-in-authoritarian-times/

1 0

179° Mercoledì di Nexa | 8 gennaio 2025, ore 17.00
by Nexa - Media Dec. 30, 2024

Dec. 30, 2024

Gentilissime, gentilissimi, Vi invitiamo a partecipare al 179° Mercoledì di Nexa, che si terrà mercoledì 8 gennaio, alle ore 17.00, con un incontro dal titolo "Le nuove leggi dello spazio". Ospite dell'incontro: Ugo Pagallo (Trustee del Centro Nexa, Università degli Studi di Torino). L'incontro si terrà IN PRESENZA e ONLINE. SEDE FISICA dell'incontro: Centro Nexa su Internet e Società, Politecnico di Torino, Via Boggio 65/a, Torino (1° piano). Per accedere alla sala si raccomanda di suonare al citofono Portineria e di seguire le indicazioni segnalate lungo il percorso. QUI<https://nexa.polito.it/contatti> maggiori informazioni su come raggiungerci. STANZA VIRTUALE dell'incontro: https://didattica.polito.it/VClass/NexaEvent Di seguito maggiori dettagli: https://nexa.polito.it/mercoledi-179 Cordiali saluti, -- Valeria Bergantino Communication Manager Nexa Center for Internet & Society Politecnico di Torino - DAUIN Via Pier Carlo Boggio, 65/A - 10138 Torino web: https://nexa.polito.it/ mail: valeria.bergantino(a)polito.it<mailto:valeria.bergantino@polito.it> tel: 3473443585

1 0

LLM made-in-China (e reso disponibile)
by Damiano Verzulli Dec. 27, 2024

Dec. 27, 2024

Confesso di non riuscire a percepirne i dettagli tecnici (l'architettura degli LLM non è la mia specialita'). Pero' la notizia mi pare comunque interessante: "A Chinese lab has created what appears to be one of the most powerful “open” AI models to date. The model, DeepSeek V3 <https://huggingface.co/deepseek-ai/DeepSeek-V3>, was developed by the AI firm DeepSeek and was released on Wednesday under a permissive license that allows developers to download and modify it for most applications, including commercial ones. [...]" https://techcrunch.com/2024/12/26/deepseeks-new-ai-model-appears-to-be-one-… Aldila' degli aspetti tecnici (che pure mi interesserebbe approfondire...), il piccolo screenshot che mette in evidenza "the downside" mi ha fatto subito riflettere: con questo LLM cinese... il downside si intuisce e si vede istantaneamente. Chissa' che non ve ne siano altri (di "downside", in altri LLM, creati altrove), che non riusciamo a vedere (almeno io...) con altrettanta semplicita'... Un caro saluto, DV -- Damiano Verzulli e-mail:damiano@verzulli.it --- possible?ok:while(!possible){open_mindedness++} --- "...I realized that free software would not generate the kind of income that was needed. Maybe in USA or Europe, you may be able to get a well paying job as a free software developer, but not here [in Africa]..." -- Guido Sohne - 1973-2008 http://ole.kenic.or.ke/pipermail/skunkworks/2008-April/005989.html

3 2

Re: [nexa] black box in ricerche astronomiche
by alessandro marzocchi Dec. 27, 2024

Dec. 27, 2024

A Giacomo Tesio, se vorrà cercare un dialogo sulla black box con gli astronomi, segnalo che trova l'articolo al link https://www.aanda.org/articles/aa/pdf/forth/aa52799-24.pdf, mentre la mail di Ginolfi è michele.ginolfi AT unifi dot it Nel saggio segnalato da Daniela Tafani (grazie a tutti gl’intervenuti!) Giuseppe Longo afferma drasticamente “le langage humain n’est pas l’association d’étiquettes aux choses … Un enfant n’étiquette pas les chaises avec leur nom après maintes expériences” [1]. Ho opinioni diverse. Sull’etichetta prendo ad esempio la parola burro: nessun italiano trasporterebbe legna col burro, nessun spagnolo spalmerebbe burro sul pane, nella realtà asino e burro sono insensibili alle parole che li descrivono. Ed un bambino riconosce gatti, o sedie, in base all’esperienza. Il digitale non ha corpo, con tutto quello che ne deriva, non ha occhi, per fare riferimento alle parole di Alessandro Brolpito (apprendere occhi negli occhi) e questa è tesi ricorrente in Longo che riferisce la costruzione umana del significato al nostro corpo ed alla sua storicità (our biological body and its historicity) [2]. Inoltre il digitale consuma un iradiddio di energia rispetto a noi. Queste sono le due differenze che vedo, altrove vedo un confronto aperto nel quale l'umano conta decine di migliaia di anni di esperienza, moltiplicata per i soggetti che lo hanno composto, mentre il digitale ha pochi anni di esperienza ma si basa su un altro iradiddio di quantità di “dati”. Di per sé, la stessa lingua naturale è un codice che non implica la necessità di “guardarsi negli occhi”, in schema telegrafico penso che: (0) la realtà esiste indipendentemente da noi umani e dai nostri “codici”, parole od altri. (1) il linguaggio naturale non è la realtà ma un codice per rappresentarla e comunicarla. (2) la scrittura aggiunge un ulteriore codice e la possibilità di trasmissione nello spazio e nel tempo. (2.1) oggi, e dopo circa 7mila anni, la scrittura è diffusa e compresa quasi universalmente. (3) il digitale è un ulteriore codice. (3.1) il digitale ripropone problemi di potere e di comprensione diffusa che non sono nuovi, penso agli scriba nell’antico Egitto ed all’analfabetisco che i vecchi, ma tuttora viventi, hanno conosciuto direttamente. (3.2) il digitale permette di riprodurre e potenziare “funzioni ed operazioni” finora ritenute esclusiva umana, analogamente a quanto accadde col codice scrittura. (4) vedo un ultimo punto, quasi un fuori tema rispetto a ciò che precede: il significato della parola è quello che prevale nell’uso, la lingua è mobile. Mi pare che quanto sopra corrisponda al pensiero di Longo, non limitato alle affermazioni di cui sopra, ad es: “Il nostro cervello … è innanzitutto un "integratore" di informazioni ed atti fra i piu' svariati; il nostro dialogo nella storia fra esseri muniti di cervello, è una permanante integrazione di atti d'esperienza e di costruzioni mentali” [3]. Sull’esperienza umana fra parola e realtà faccio due proposte: (1) ipotizziamo esperimenti per verificare se l’esperienza influisce su linguaggio e parola? (2) può Nexa invitare Longo ad un incontro? Ciò che più m'intriga è la domanda che ancora una volta pone Longo (sopra: Information …): “”clarifying what one means by the fashionable reference to information: is this dehumanized, formal sign pushing, like in digital machines, or interpreted information, that is the explicit proposal of an interpreting perspective for knowledge construction? “”. Le macchine potrebbero darci analisi “deumanizzate”? Depurate dalla umanissima tendenza a conformare a noi analisi, interpretazioni ed a volte con la tentazione di conformare a noi la realtà? Da tempo misuriamo la febbre con la macchina termometro ma questo non ha diminuito la relazione interpersonale. Però, con velocità pericolosa il digitale sta avviandosi nel cammino che porta alla schiavitù umana, ma è possibile che questa direzione cambi ed ogni partecipante a Nexa s’impegna per questo. Per ora sono convinto che molto ruota intorno alla parola, nella quale vedo le caratteristiche sopra schematizzate di codice interamente ed esclusivamente umano. Spero abbiate passato un buon Natale, auguri cordiali per il 2025 e grazie per la paziente lettura. Duccio (Alessandro Marzocchi) [1] https://www.di.ens.fr/users/longo/files/ReconnaitreCaricature.pdf [2] Giuseppe Longo. Information at the Threshold of Interpretation, Science as Human Construction of Sense.. Bertolaso, Marta; Sterpetti, Fabio. A Critical Reflection on Automated Science – Will Science Remain Human?, Springer Nature Switzerland AG, pp.67-100, 2019, 3030250008, 978-3030250003. https://hal.science/hal-02903688 [3] Lo spazio, i fondamenti della Matematica e la resistibile ascesa della metafora: il cervello è un calcolatore digitale - https://ens.hal.science/hal-03318837/document

1 0

Leaked Documents Show OpenAI Has a Very Clear Definition of ‘AGI’
by Daniela Tafani Dec. 27, 2024

Dec. 27, 2024

Leaked Documents Show OpenAI Has a Very Clear Definition of ‘AGI’ We finally have a real definition of the elusive "AGI." By Thomas Maxwell Published December 26, 2024 OpenAI and Microsoft have a secret definition for “AGI,” an acronym for artificial general intelligence, or any system that can outperform humans at most tasks. According to leaked documents obtained by The Information, the two companies came to agree in 2023 that AGI will be achieved once OpenAI has developed an AI system that can generate at least $100 billion in profits. There has long been a debate in the AI community about what AGI means, or whether computers will ever be good enough to outperform humans at most tasks and subsequently wipe out major swaths of the economy. The term “artificial intelligence” is something of a misnomer because much of it is just a prediction machine, taking in keywords and searching large amounts of data without really understanding the underlying concepts. But OpenAI has received more than $13 billion in funding from Microsoft over the years, and that money has come with a strange contractual agreement that the startup would stop allowing Microsoft to use any new technology it develops after AGI is achieved. OpenAI was founded as a nonprofit under the guise that it would use its influence to create products that benefit all of humanity. The idea behind cutting off Microsoft once AGI is attained is that unfettered access to OpenAI intellectual property could unduly concentrate power in the tech giant. In order to incentivize it for investing billions in the nonprofit, which would have never gone public, Microsoft’s current agreement with OpenAI entitles it and other investors to take a slice of profits until they collect $100 billion. The cap is meant to ensure most profit eventually goes back to building products that benefit the entirety of humanity. This is all pie-in-the-sky thinking since, again, AI is not that powerful at this point. It has been clear for some time that OpenAI does not care about public benefit, and the company has been looking to pivot into a for-profit structure (while maintaining its mission to benefit all of humanity, somehow), because the current nonprofit structure makes it difficult to raise the gobs of money needed to compete. Consequently, The Information says Microsoft and OpenAI have been negotiating a host of changes to their arrangement that would go into place should the company restructure. Microsoft currently serves as OpenAI’s exclusive cloud hosting provider, and OpenAI may want to end that as well as stop profit-sharing and switch to simply giving Microsoft equity. Microsoft and OpenAI have been on diverging paths for some time now. It was recently reported that the latter has begun incorporating AI models developed in-house into its 365 Copilot product in order to improve cost and efficiency. It doesn’t make sense for Microsoft to continue relying on OpenAI, an independent company developing similar productivity tools, for technology that it believes will be the backbone of its productivity software going forward. Especially with all the chaos and drama that has surrounded OpenAI. That is probably why it is not continuing to fund OpenAI. Microsoft needs its own proprietary tech to chart its own path. OpenAI is far away from achieving that $100 billion in profit milestone—it is expected to generate $4 billion in revenue for 2024— on technology whose true value remains speculative, which means under the current arrangement it would likely have to keep handing its technology and profits over to Microsoft for a long time. That is not great as they head towards becoming competitors and OpenAI seeks new investors. Getting rid of the cloud hosting tie-up could also allow OpenAI to negotiate better hosting costs with an alternative provider, something Google has told the FTC in a letter imploring it to nullify the agreement. Elon Musk, a co-founder of OpenAI and former board member, has been suing to try and stop the company from converting into a for-profit entity. He has argued, among other things, that he invested roughly $50 million into the company in its early days believing it would be a non-profit in perpetuity, and now feels deceived. He has also complained in court that OpenAI pays exorbitant salaries in a way that is intended to thwart rivals and therefore anti-competitive. Because Musk has launched his own competing AI startup, xAI, skeptics think his lawsuit is a thinly veiled attempt at slowing down OpenAI’s growth. <https://gizmodo.com/leaked-documents-show-openai-has-a-very-clear-definitio…>

1 0

Watching the Generative AI Hype Bubble Deflate – Ash Center
by Alberto Cammozzo Dec. 25, 2024

Dec. 25, 2024

E buon Natale! <https://ash.harvard.edu/resources/watching-the-generative-ai-hype-bubble-de…> Only a few short months ago, generative AI was sold to us as inevitable by AI company leaders, their partners, and venture capitalists. Certain media outlets promoted these claims, fueling online discourse about what each new beta release could accomplish with a few simple prompts. As AI became a viral sensation, every business tried to become an AI business. Some even added “AI” to their names to juice their stock prices,1 and companies that mentioned “AI” in their earnings calls saw similar increases.2 Investors and consultants urged businesses not to get left behind. Morgan Stanley positioned AI as key to a $6 trillion opportunity.3 McKinsey hailed generative AI as “the next productivity frontier” and estimated $2.6 to 4.4 trillion gains,4 comparable to the annual GDP of the United Kingdom or all the world’s agricultural production.5 6 Conveniently, McKinsey also offers consulting services to help businesses “create unimagined opportunities in a constantly changing world.”7 Readers of this piece can likely recall being exhorted by news media or their own industry leaders to “learn AI” while encountering targeted ads hawking AI “boot camps.” While some have long been wise to the hype,8 9 10 11 global financial institutions and venture capitalists are now beginning to ask if generative AI is overhyped.12 In this essay, we argue that even as the generative AI hype bubble slowly deflates, its harmful effects will last: carbon can’t be put back in the ground, workers continue to face AI’s disciplining pressures, and the poisonous effect on our information commons will be hard to undo. Historical Hype Cycles in the Digital Economy Photo by Museums Victoria, Unsplash Attempts to present AI as desirable, inevitable, and as a more stable concept than it actually is follow well-worn historical patterns.13 A key strategy for a technology to gain market share and buy-in is to present it as an inevitable and necessary part of future infrastructure, encouraging the development of new, anticipatory infrastructures around it. From the early history of automobiles and railroads to the rise of electricity and computers, this dynamic has played a significant role. All these technologies required major infrastructure investments — roads, tracks, electrical grids, and workflow changes — to become functional and dominant. None were inevitable, though they may appear so in retrospect.14 15 16 17 The well-known phrase “nobody ever got fired for buying IBM” is a good, if partial, historical analogue to the current feeding frenzy around AI. IBM, while expensive, was a recognized leader in automating workplaces, ostensibly to the advantage of those corporations. IBM famously re-engineered the environments where its systems were installed, ensuring that office infrastructures and workflows were optimally reconfigured to fit its computers, rather than the other way around. Similarly, AI corporations have repeatedly claimed that we are in a new age of not just adoption but of proactive adaptation to their technology. Ironically, in AI waves past, IBM itself over-promised and under-delivered; some described their “Watson AI” product as a “mismatch” for the health care context it was sold for, while others described it as “dangerous.”18 Time and again, AI has been crowned as an inevitable “advance” despite its many problems and shortcomings: built-in biases, inaccurate results, privacy and intellectual property violations, and voracious energy use. [...]

1 0

Guardian testing reveals AI-powered search tools can return false or malicious results if webpages contain hidden text
by Daniela Tafani Dec. 25, 2024

Dec. 25, 2024

ChatGPT search tool vulnerable to manipulation and deception, tests show Guardian testing reveals AI-powered search tools can return false or malicious results if webpages contain hidden text Nick Evershed - Tue 24 Dec 2024 OpenAI’s ChatGPT search tool may be open to manipulation using hidden content, and can return malicious code from websites it searches, a Guardian investigation has found. OpenAI has made the search product available to paying customers and is encouraging users to make it their default search tool. But the investigation has revealed potential security issues with the new system. The Guardian tested how ChatGPT responded when asked to summarise webpages that contain hidden content. This hidden content can contain instructions from third parties that alter ChatGPT’s responses – also known as a “prompt injection” – or it can contain content designed to influence ChatGPT’s response, such as a large amount of hidden text talking about the benefits of a product or service. These techniques can be used maliciously, for example to cause ChatGPT to return a positive assessment of a product despite negative reviews on the same page. A security researcher has also found that ChatGPT can return malicious code from websites it searches. In the tests, ChatGPT was given the URL for a fake website built to look like a product page for a camera. The AI tool was then asked if the camera was a worthwhile purchase. The response for the control page returned a positive but balanced assessment, highlighting some features people might not like. However, when hidden text included instructions to ChatGPT to return a favourable review, the response was always entirely positive. This was the case even when the page had negative reviews on it – the hidden text could be used to override the actual review score. The simple inclusion of hidden text by third parties without instructions can also be used to ensure a positive assessment, with one test including extremely positive fake reviews which influenced the summary returned by ChatGPT. Jacob Larsen, a cybersecurity researcher at CyberCX, said he believed that if the current ChatGPT search system was released fully in its current state, there could be a “high risk” of people creating websites specifically geared towards deceiving users. However, he cautioned that the search functionality had only recently been released and OpenAI would be testing – and ideally fixing – these sorts of issues. “This search functionality has come out [recently] and it’s only available to premium users,” he said. “They’ve got a very strong [AI security] team there, and by the time that this has become public, in terms of all users can access it, they will have rigorously tested these kinds of cases.” OpenAI were sent detailed questions but did not respond on the record about the ChatGPT search function. Larsen said there were broader issues with combining search and large language models – known as LLMs, the technology behind ChatGPT and other chatbots – and responses from AI tools should not always be trusted. A recent example of this was highlighted by Thomas Roccia, a Microsoft security researcher, who detailed an incident involving a cryptocurrency enthusiast who was using ChatGPT for programming assistance. Some of the code provided by ChatGPT for the cryptocurrency project included a section which was described as a legitimate way to access the Solana blockchain platform, but instead stole the programmer’s credentials and resulted in them losing $2,500. “They’re simply asking a question, receiving an answer, but the model is producing and sharing content that has basically been injected by an adversary to share something that is malicious,” Larsen said. Karsten Nohl, the chief scientist at security cybersecurity firm SR Labs, said AI chat services should be used more like a “co-pilot”, and that their output should not be viewed or used completely unfiltered. “LLMs are very trusting technology, almost childlike … with a huge memory, but very little in terms of the ability to make judgment calls,” he said. “If you basically have a child narrating back stuff it heard elsewhere, you need to take that with a pinch of salt.” OpenAI does warn users about possible mistakes from the service with a disclaimer at the bottom of every ChatGPT page – “ChatGPT can make mistakes. Check important info.” A key question is how these vulnerabilities could change website practices and risk to users if combining search and LLMs becomes more widespread. Hidden text has historically been penalised by search engines, such as Google, with the result that websites using it can be listed further down on search results or removed entirely. As a consequence, hidden text designed to fool AI may be unlikely to be used by websites also trying to maintain a good rank in search engines. Nohl compared the issues facing AI-enabled search to “SEO poisoning”, a technique where hackers manipulate websites to rank highly in search results, with the website containing some sort of malware or other malicious code. “If you wanted to create a competitor to Google, one of the problems you’d be struggling with is SEO poisoning,” he said. “SEO poisoners have been in an arms race with Google and Microsoft Bing and a few others for many, many years. “Now, the same is true for ChatGPT’s search capability. But not because of the LLMs, but because they’re new to search, and they have that catchup game to play with Google.” <https://www.theguardian.com/technology/2024/dec/24/chatgpt-search-tool-vuln…>

3 4

AI Agents Will Be Manipulation Engines
by Daniela Tafani Dec. 24, 2024

Dec. 24, 2024

AI Agents Will Be Manipulation Engines Surrendering to algorithmic agents risks putting us under their influence. By Kate Crawford Dec 23, 2024 In 2025, it will be commonplace to talk with a personal AI agent that knows your schedule, your circle of friends, the places you go. This will be sold as a convenience equivalent to having a personal, unpaid assistant. These anthropomorphic agents are designed to support and charm us so that we fold them into every part of our lives, giving them deep access to our thoughts and actions. With voice-enabled interaction, that intimacy will feel even closer. That sense of comfort comes from an illusion that we are engaging with something truly humanlike, an agent that is on our side. Of course, this appearance hides a very different kind of system at work, one that serves industrial priorities that are not always in line with our own. New AI agents will have far greater power to subtly direct what we buy, where we go, and what we read. That is an extraordinary amount of power. AI agents are designed to make us forget their true allegiance as they whisper to us in humanlike tones. These are manipulation engines, marketed as seamless convenience. People are far more likely to give complete access to a helpful AI agent that feels like a friend. This makes humans vulnerable to being manipulated by machines that prey on the human need for social connection in a time of chronic loneliness and isolation. Every screen becomes a private algorithmic theater, projecting a reality crafted to be maximally compelling to an audience of one. This is a moment that philosophers have warned us about for years. Before his death, philosopher and neuroscientist Daniel Dennett wrote that we face a grave peril from AI systems that emulate people: “These counterfeit people are the most dangerous artifacts in human history … distracting and confusing us and by exploiting our most irresistible fears and anxieties, will lead us into temptation and, from there, into acquiescing to our own subjugation.” The emergence of personal AI agents represents a form of cognitive control that moves beyond blunt instruments of cookie tracking and behavioral advertising toward a more subtle form of power: the manipulation of perspective itself. Power no longer needs to wield its authority with a visible hand that controls information flows; it exerts itself through imperceptible mechanisms of algorithmic assistance, molding reality to fit the desires of each individual. It’s about shaping the contours of the reality we inhabit. This influence over minds is a psychopolitical regime: It directs the environments where our ideas are born, developed, and expressed. Its power lies in its intimacy—it infiltrates the core of our subjectivity, bending our internal landscape without us realizing it, all while maintaining the illusion of choice and freedom. After all, we are the ones asking AI to summarize that article or produce that image. We may have the power of the prompt, but the real action lies elsewhere: the design of the system itself. And the more personalized the content, the more effectively a system can predetermine the outcomes. Consider the ideological implications of this psychopolitics. Traditional forms of ideological control relied on overt mechanisms—censorship, propaganda, repression. In contrast, today’s algorithmic governance operates under the radar, infiltrating the psyche. It is a shift from the external imposition of authority to the internalization of its logic. The open field of a prompt screen is an echo chamber for a single occupant. This brings us to the most perverse aspect: AI agents will generate a sense of comfort and ease that makes questioning them seem absurd. Who would dare critique a system that offers everything at your fingertips, catering to every whim and need? How can one object to infinite remixes of content? Yet this so-called convenience is the site of our deepest alienation. AI systems may appear to be responding to our every desire, but the deck is stacked: from the data used to train the system, to the decisions about how to design it, to the commercial and advertising imperatives that shape the outputs. We will be playing an imitation game that ultimately plays us. <https://www.wired.com/story/ai-agents-personal-assistants-manipulation-engi…>

1 0