AI, Kurapati (Ceo di Clearbox AI): ‘Dati sintetici per aiutare le aziende’

Aiutare le aziende ad usare al meglio l’Intelligenza Artificiale, in modo semplice e corretto tramite i dati sintetici. Questa in sintesi la mission di Clearbox AI, una startup torinese specializzata in AI, la cui Ceo Shalini Kurapati è stata premiata dalla Commissione Ue come una delle principali ‘Lady Tech’ europee. Kurapati prenderà parte all’evento ‘State of Privacy ‘23’ organizzato dal Garante Privacy il 18 settembre a Roma, per parlare dell’impatto dell’AI sulla società.

Key4biz. Clearbox AI, start up di intelligenza artificiale, per fare cosa?

Shalini Kurapati. Clearbox AI è una startup tech italiana che aiuta le aziende a lanciare progetti di Intelligenza Artificiale attraverso l’utilizzo di Dati Sintetici. Nasciamo nel 2019 col supporto dell’Incubatore del Politecnico di Torino, abbiamo fondato la società con 4 co-fondatori, me compresa, che si sono uniti dopo una lunga esperienza di ricerca e professionale in molti paesi europei. La nostra missione: comprendere gli ostacoli che le aziende incontrano nello sviluppo di Intelligenza Artificiale.

Key4biz. Quali sono i principali ostacoli per le aziende?

Shalini Kurapati. Abbiamo scoperto che molto spesso queste sfide sono legate ai dati, quando questi sono sensibili e difficili da gestire a livello di privacy, la loro quantità non è sufficiente, o non sono abbastanza rappresentativi per tutte le fasce di popolazione per garantire risultati di successo.

Key4biz. E’ qui che entrano in gioco i Dati Sintetici?

Shalini Kurapati. Sì. I Dati Sintetici possono risolvere questi problemi. Sono generati artificialmente da algoritmi AI sulla base dei dati originali, dei quali mantengono le proprietà statistiche e il potere predittivo, risultando quindi realistici. Essendo simili ma non uguali ai dati reali, non contengono informazioni personali e possono essere condivisi rispettando i regolamenti di privacy, come il GDPR. Offriamo un prodotto/software basato sulla nostra tecnologia proprietaria che genera dati sintetici di alta qualità per le aziende che vogliono iniziare e accelerare il loro percorso di Intelligenza artificiale. Il nostro team ha una duplice anima: forti radici nel mondo R&D con lauree dalle migliori università tecnologiche in Europa, che vanno mano nella mano con un approccio ingegneristico e pratico nel risolvere i problemi dei clienti. Le strade professionali e personali dei componenti del gruppo hanno attraversato l’Europa per convergere a Torino, dove è nata la missione Clearbox AI: creare valore per le imprese attraverso le tecnologie IA, rispettandone i principi etici.

Key4biz. Ci può fare qualche esempio concreto di come i Dati Sintetici possono aiutare le aziende?

Shalini Kurapati. Oggi avere un modello di Intelligenza Artificiale è molto facile. Basta un download. Ma la sfida vera è avere dati giusti dal punto di vista dell’accesso privacy e della qualità. I dati sintetici sono una delle soluzioni a questo problema. Sono generati artificialmente tramite algoritmi. Noi produciamo dati sintetici per due motivi principali: la tutela della privacy e poi in molti casi i data set non sono rappresentativi per diversi motivi, il che può creare pregiudizi (bias) dell’algoritmo. Questo perché l’algoritmo pensa erroneamente che questo set di dati rappresenti la realtà, ma non è così. I dati sintetici possono aumentare questo data set sbilanciato in modo tale da far rappresentare la realtà.  

Key4biz. Ad esempio?

Shalini Kurapati. Ad esempio, sviluppiamo un algoritmo per riconoscere un tumore sulla pelle. L’algoritmo deve capire cosa vuol dire un algoritmo sulla pelle. Noi forniamo diversi esempi di tumori. Ma se forniamo soltanto esempi con carnagione bianca questo algoritmo non funziona per altre carnagioni di pelle. Quindi funziona soltanto per questo tipo di pelle. Un dato non rappresentativo può danneggiare altre persone se usiamo questo algoritmo senza testare tutti i tipi di persone.     

Key4biz. E ancora?

Shalini Kurapati. Storicamente le donne non erano tantissime nel mondo del lavoro. Quindi, se alleniamo un algoritmo soltanto con dati storici, se anche mettiamo in produzione con dati attuali, l’algoritmo pensa che i dati storici sono la realtà. Magari prende delle decisioni sulla base del fatto che 30 anni fa la percentuale di donne al lavoro erano meno e quindi prende decisioni in base a dati ormai superati. Qui vengono in soccorso i dati sintetici, che prendono sempre in considerazione la realtà.

Key4biz. Un altro esempio dal campo medico?

Shalini Kurapati. Per tanto tempo i dati della ricerca sono sempre stati presi su degli uomini. Per tante malattie i sintomi delle donne non sono presenti e sono troppo pochi. Per esempio, per un infarto i sintomi di un uomo sono male al cuore e al braccio. Per una donna, invece, può essere mal di schiena. Però, se alleniamo un modello è successo anche questo, che un’azienda che ha sviluppato un algoritmo di intelligenza artificiale per riconoscere una malattia dai sintomi, per un uomo ha detto di andare subito al pronto soccorso. Ma per una donna ha detto che era un attacco di panico, calmati! Invece era un infarto. Quando parliamo di responsabilità, un algoritmo soprattutto per una applicazione per prendere decisioni su un curriculum o prendere decisioni su un mutuo, o decisioni sulla salute, quindi un impatto grande, dobbiamo avere una metodologia per capire quali dati stiamo usando. Come possiamo migliorare questo. Lo stesso discorso vale anche per il riconoscimento di frodi per banche e assicurazioni. Qui il problema è che gli esempi di frodi non sono molto numerosi ed è quindi difficile per l’algoritmo riconoscerle. Grazie ai dati sintetici possiamo fornire all’algoritmo più esempi per riconoscere le frodi.

Key4biz. Quali sono le aree di competenza del team?

Shalini Kurapati. Le aree di competenza del team spaziano tra machine learning, data science, data privacy, uncertainty quantification, project management, innovazione e sviluppo software, e sono state fondamentali per creare una visione condivisa.

Key4biz. La Commissione Europea l’ha premiata come una delle 50 più geniali e innovative ‘Lady Tech’. Per questo motivo è entrata nel progetto pilota, voluto da Bruxelles, per supportare start up ad elevato contenuto tecnologico guidate da donne. Come considera l’approccio di Bruxelles sull’IA? (vedi AI ACT)

Shalini Kurapati. Sì, siamo molto contenti di essere stati selezionati tra i vincitori del progetto Women TechEU, finanziato dalla Commissione Europea. L’iniziativa ha come obiettivo il supporto alle startup deep-tech guidate da donne per valorizzare il talento e favorire un’innovazione più inclusiva nell’ecosistema tech europeo. In questo senso, i dati sintetici che produciamo giocano un ruolo fondamentale per gli obiettivi del bando, per i motivi di mitigazione bias che abbiamo raccontato. Anche la nostra missione aziendale è in linea con questi fini perché si colloca all’intersezione tra tecnologia ed etica dell’Intelligenza Artificiale. In particolare, utilizzeremo i finanziamenti della Commissione Europea per creare un modulo basato sui dati sintetici che punta a diminuire la presenza di bias nei progetti di Intelligenza Artificiale. Credo che a livello europeo si stiano facendo molti passi avanti nella promozione della conoscenza e creazione di framework legati all’AI. Uno su tutti è l’AI Act, apripista nell’istituzione di regole esaustive e realmente applicabili per garantire la sicurezza dell’IA e l’innovazione responsabile, attraverso un approccio basato sul rischio delle applicazioni IA. Nonostante alcuni limiti (se ne è parlato per anni e ci sono sempre discussioni in corso) e sebbene potrebbe richiedere da 2 a 3 anni per entrare in vigore, ci sono comunque molte ragioni per sostenerlo fortemente. Ha creato un punto di riferimento e un quadro per la cooperazione nella regolamentazione globale dell’IA. Ha attivato opportunità per adottare codici di condotta e patti sull’IA per iniziare a mettere in pratica i principi della regolamentazione mentre aspettiamo che entrino in vigore e per cominciare ad adattarsi all’evoluzione della tecnologia IA. A complicare il tentativo dell’Ue, però, è anche la portata di una norma che, per essere efficace, deve essere riconosciuta a livello globale. Altrimenti rischia di essere agevolmente aggirata. Infatti le principali società che sviluppano l’AI nel mondo sono oggi Microsoft/Open AI, Google, Meta, Tesla, solo per citarne alcune: un panorama che coinvolge gli Stati Uniti, ma che inizia a coinvolgere anche la Cina. E gli Stati Uniti, per il momento, non hanno alcuna intenzione di frenare la ricerca e lo sviluppo: per quanto l’amministrazione Biden sia consapevole dei rischi, dopo tanti discorsi tutto ciò che è stato fatto ad oggi – di fatto – sono una serie di misure non legislative che suggeriscono ai colossi dell’AI come dovrebbero comportarsi per auto-regolare la propria tecnologia. Dobbiamo tutti essere molto interessati e proattivi nell’implementare i requisiti dell’AI Act, non solo per essere conformi, ma per garantirci un futuro responsabile!

Key4biz. Come governare i rischi dell’AI e come l’Italia, secondo lei, dovrebbe sostenere di più lo sviluppo dell’AI in Italia?

Shalini Kurapati. Spesso si parla dell’Italia come terra di cultura e di buon cibo, il che è assolutamente vero, ma dovremmo considerare che l’Italia ha già una lunga storia e saper-fare nell’industrializzazione e soprattutto nell’automazione industriale. In termini di robot density (un’importante indicatore dell’automazione industriale), l’Italia è davanti a Francia e Canada. Tuttavia, anche se non si può dire lo stesso dell’adozione delle applicazioni di IA e dell’innovazione guidata dai dati, l’Italia ha un grande potenziale. Abbiamo bisogno di una bella spinta nella giusta direzione in termini di creazione di maggiore consapevolezza e di cambiamento di mentalità. Ad oggi siamo molto felici di riscontrare sempre più interesse da parte delle aziende italiane nell’approfondire come si può implementare l’AI in maniera efficiente all’interno dei processi e, da parte nostra, cerchiamo da sempre di diffondere messaggi precisi per aiutarle a beneficiarne. Ad esempio, che l’AI non è appannaggio solo delle grandi aziende. Allo stesso tempo, però, bisogna avere una conoscenza precisa degli obiettivi per i quali si vuole usare l’AI, affidarsi a delle professionalità specifiche e ridimensionare le aspettative dell’AI come “panacea di tutti i mali”.

“State of Privacy ’23”, il 18 settembre l’evento del Garante. Scopri il programma

https://www.key4biz.it/ai-kurapati-ceo-di-clearbox-ai-dati-sintetici-per-aiutare-le-aziende/459453/




Private AI summit with senators, titans of tech garners controversy

WASHINGTON, DC - SEPTEMBER 13: Google CEO Sundar Pichai and Meta CEO Mark Zuckerberg arrive to the Senate bipartisan Artificial Intelligence (AI) Insight Forum on Capitol Hill in Washington, DC, on September 13, 2023. (Photo by Elizabeth Frantz for The Washington Post)
Enlarge / Google CEO Sundar Pichai and Meta CEO Mark Zuckerberg arrive to the Senate bipartisan Artificial Intelligence (AI) Insight Forum on Capitol Hill in Washington, DC, on September 13, 2023.
Getty Images

On Wednesday, US Senator Chuck Schumer (D-N.Y.) hosted an “AI Insight Forum” in the Senate’s office building about potential AI regulation. Attendees included billionaires and modern-day industry titans such as Elon Musk, Bill Gates, Mark Zuckerberg, OpenAI’s Sam Altman, and Jensen Huang of Nvidia. But this heavily corporate guest list—with 14 out of 22 being CEOs—had some scratching their heads.

“This is the room you pull together when your staffers want pictures with tech industry AI celebrities. It’s not the room you’d assemble when you want to better understand what AI is, how (and for whom) it functions, and what to do about it,” wrote Signal President Meredith Whittaker on X.

The CEO-heavy list had others questioning the technical AI acumen of the attendees. “I hope that Schumer and others pivot to hear counterpoints from the many technical people who are good at explaining tech,” said Dr. Margaret Mitchell of AI platform Hugging Face, whose CEO attended the meeting. “These CEOs are largely not. And they are very much incentivized to obscure critical details, to the extent they themselves understand them.”

Mitchell also raised concerns about the lack of technically skilled women in the meeting. Overall, seven of the 22 invitees were women, but not all of them hold technical roles in AI, including representatives from the AFL-CIO and the Writers Guild. “Women are technical, but you wouldn’t know that from Schumer’s invitees—and this is a massive problem if we want to create AI regulation that helps AI work for everyone,” Mitchell said.

Elon Musk, CEO of Tesla and X, speaks to reporters as he leaves the “AI Insight Forum” at the Russell Senate Office Building on Capitol Hill on September 13, 2023, in Washington, DC.
Enlarge / Elon Musk, CEO of Tesla and X, speaks to reporters as he leaves the “AI Insight Forum” at the Russell Senate Office Building on Capitol Hill on September 13, 2023, in Washington, DC.
Getty Images

Some Senators criticized the nature of the meeting as well. “I think it’s ridiculous that all these monopolists are all here to tell senators how to shape the regulatory framework so they can make the maximum amount of money,” said Senator Josh Hawley (R-Mo.).

Despite the controversy, the meeting drew bipartisan interest, with more than 60 senators participating. Democratic Senate Majority Leader Schumer and Senator Todd Young (R-Ind.) expressed optimism about the Senate’s readiness to consider legislative proposals. However, Senator Mike Rounds (R-S.D.) pointed out that the process of drafting appropriate legislation would take time. “Are we ready to go out and write legislation? Absolutely not,” Reuters quoted Rounds as saying. “We’re not there.”

While the meeting was closed to media and the public, Reuters and The New York Times reported on some of the discussion, conducted in private so that no one would “play to the press,” Schumer said.

According to Reuters, Musk argued for the necessity of a “referee” in the AI space. He described the regulation of AI as akin to a sports game, where rules ensure fair play and protect the participants. “It’s important for us to have a referee to ensure that companies take actions that are safe and in the interest of the general public,” Musk said.

The multi-billionaire, who manages various tech ventures, including the social media platform X, framed the meeting as a “service to humanity,” suggesting its ramifications “may go down in history as very important to the future of civilization.” Musk’s comments echo his previous calls for a six-month pause on the development of AI systems more powerful than OpenAI’s GPT-4, which some critics say overlooked current harm from AI in favor of unproven, hypothetical dangers.

On the other hand, Meta’s Zuckerberg reportedly pushed for a collaborative approach between the government and tech companies. He urged Congress to “engage with AI to support innovation and safeguards,” highlighting the strategic advantage of having American companies set global standards.

Due to the rapid rise of generative AI in the public eye over the past year—especially related to the success of AI assistant ChatGPT and warnings about potential dangers of hypothetical super-intelligent machines—artificial intelligence tools have fallen squarely into the crosshairs of potential US government regulation, with heavy interest recently from Schumer and others in Congress. Last week, US Senators Richard Blumenthal (D-Conn.) and Hawley proposed a bipartisan legislative framework that would require companies working on “high-risk” AI applications, such as GPT-4, to obtain a government license.

The call for AI regulation has become increasingly global, with countries worldwide grappling to set rules about facial recognition, deepfakes, training data sets, and more. This week, companies Adobe, IBM, and Nvidia announced they had joined President Joe Biden’s voluntary AI commitments, which mandate measures like watermarking AI-generated content to flag its artificial origins.

With so much hype out there—and so many dollars flying around—industry experts like Mitchell are worried that the technical details may get lost in the conversation. “The right thing to do would have been to communicate first with people who are technically skilled and who can speak to people’s rights,” said Mitchell. “They would have set the foundation of Schumer’s understanding. Instead the foundation of his understanding is now aligned to short-term profit maximization.”

https://arstechnica.com/?p=1968348




Tech Industry Leaders Endorse Regulating Artificial Intelligence at Rare Summit in Washington

The nation’s biggest technology executives on Wednesday loosely endorsed the idea of government regulations for artificial intelligence at an unusual closed-door meeting in the U.S. Senate. But there is little consensus on what regulation would look like, and the political path for legislation is difficult.

Senate Majority Leader Chuck Schumer, who organized the private forum on Capitol Hill as part of a push to legislate artificial intelligence, said he asked everyone in the room — including almost two dozen tech executives, advocates and skeptics — whether government should have a role in the oversight of artificial intelligence, and “every single person raised their hands, even though they had diverse views,” he said.

Among the ideas discussed was whether there should be an independent agency to oversee certain aspects of the rapidly-developing technology, how companies could be more transparent and how the United States can stay ahead of China and other countries.

“The key point was really that it’s important for us to have a referee,” said Elon Musk, CEO of Tesla and X, during a break in the daylong forum. “It was a very civilized discussion, actually, among some of the smartest people in the world.”

Schumer will not necessarily take the tech executives’ advice as he works with colleagues on the politically difficult task of ensuring some oversight of the burgeoning sector. But he invited them to the meeting in hopes that they would give senators some realistic direction for meaningful regulation.

Congress should do what it can to maximize AI’s benefits and minimize the negatives, Schumer said, “whether that’s enshrining bias, or the loss of jobs, or even the kind of doomsday scenarios that were mentioned in the room. And only government can be there to put in guardrails.”

Other executives attending the meeting were Meta’s Mark Zuckerberg, former Microsoft CEO Bill Gates and Google CEO Sundar Pichai. Musk said the meeting “might go down in history as being very important for the future of civilization.”

Advertisement. Scroll to continue reading.

First, though, lawmakers have to agree on whether to regulate, and how.

Learn More at SecurityWeek’s Cyber AI & Automation Summit

Join this virtual event as we explore the hype and promise surrounding AI-powered security solutions in the enterprise and the threats posed by adversarial use of AI.

Cyber AI & Automation Summit

December 6, 2023 | Virtual Event

Congress has a lackluster track record when it comes to regulating new technology, and the industry has grown mostly unchecked by government in the past several decades. Many lawmakers point to the failure to pass any legislation surrounding social media, such as for stricter privacy standards.

Schumer, who has made AI one of his top issues as leader, said regulation of artificial intelligence will be “one of the most difficult issues we can ever take on,” and he listed some of the reasons why: It’s technically complicated, it keeps changing and it “has such a wide, broad effect across the whole world,” he said.

Sparked by the release of ChatGPT less than a year ago, businesses have been clamoring to apply new generative AI tools that can compose human-like passages of text, program computer code and create novel images, audio and video. The hype over such tools has accelerated worries over its potential societal harms and prompted calls for more transparency in how the data behind the new products is collected and used.

Republican Sen. Mike Rounds of South Dakota, who led the meeting with Schumer, said Congress needs to get ahead of fast-moving AI by making sure it continues to develop “on the positive side” while also taking care of potential issues surrounding data transparency and privacy.

“AI is not going away, and it can do some really good things or it can be a real challenge,” Rounds said.

The tech leaders and others outlined their views at the meeting, with each participant getting three minutes to speak on a topic of their choosing. Schumer and Rounds then led a group discussion.

During the discussion, according to attendees who spoke about it, Musk and former Google CEO Eric Schmidt raised existential risks posed by AI, and Zuckerberg brought up the question of closed vs. “open source” AI models. Gates talked about feeding the hungry. IBM CEO Arvind Krishna expressed opposition to proposals favored by other companies that would require licenses.

In terms of a potential new agency for regulation, “that is one of the biggest questions we have to answer and that we will continue to discuss,” Schumer said. Musk said afterward he thinks the creation of a regulatory agency is likely.

Outside the meeting, Google CEO Pichai declined to give details about specifics but generally endorsed the idea of Washington involvement.

“I think it’s important that government plays a role, both on the innovation side and building the right safeguards, and I thought it was a productive discussion,” he said.

Some senators were critical that the public was shut out of the meeting, arguing that the tech executives should testify in public.

Sen. Josh Hawley, R-Mo., said he would not attend what he said was a “giant cocktail party for big tech.” Hawley has introduced legislation with Sen. Richard Blumenthal, D-Conn., to require tech companies to seek licenses for high-risk AI systems.

“I don’t know why we would invite all the biggest monopolists in the world to come and give Congress tips on how to help them make more money and then close it to the public,” Hawley said.

While civil rights and labor groups were also represented at the meeting, some experts worried that Schumer’s event risked emphasizing the concerns of big firms over everyone else.

Sarah Myers West, managing director of the nonprofit AI Now Institute, estimated that the combined net worth of the room Wednesday was $550 billion and it was “hard to envision a room like that in any way meaningfully representing the interests of the broader public.” She did not attend.

In the United States, major tech companies have expressed support for AI regulations, though they don’t necessarily agree on what that means. Similarly, members of Congress agree that legislation is needed, but there is little consensus on what to do.

There is also division, with some members of Congress worrying more about overregulation of the industry while others are concerned more about the potential risks. Those differences often fall along party lines.

“I am involved in this process in large measure to ensure that we act, but we don’t act more boldly or over-broadly than the circumstances require,” Young said. “We should be skeptical of government, which is why I think it’s important that you got Republicans at the table.”

Some concrete proposals have already been introduced, including legislation by Sen. Amy Klobuchar, D-Minn., that would require disclaimers for AI-generated election ads with deceptive imagery and sounds. Schumer said they discussed “the need to do something fairly immediate” before next year’s presidential election.

Hawley and Blumenthal’s broader approach would create a government oversight authority with the power to audit certain AI systems for harms before granting a license.

Some of those invited to Capitol Hill, such as Musk, have voiced dire concerns evoking popular science fiction about the possibility of humanity losing control to advanced AI systems if the right safeguards are not in place. But the only academic invited to the forum, Deborah Raji, a University of California, Berkeley researcher who has studied algorithmic bias, said she tried to emphasize real-world harms already occurring.

“There was a lot of care to make sure the room was a balanced conversation, or as balanced as it could be” Raji said. What remains to be seen, she said, is which voices senators will listen to and what priorities they elevate as they work to pass new laws.

Some Republicans have been wary of following the path of the European Union, which signed off in June on the world’s first set of comprehensive rules for artificial intelligence. The EU’s AI Act will govern any product or service that uses an AI system and classify them according to four levels of risk, from minimal to unacceptable.

A group of European corporations has called on EU leaders to rethink the rules, arguing that it could make it harder for companies in the 27-nation bloc to compete with rivals overseas in the use of generative AI.

Related: The Good, the Bad and the Ugly of Generative AI

Related: ChatGPT, the AI Revolution, and Security, Privacy and Ethical Implications

Related: XDR and the Age-old Problem of Alert Fatigue

Related: House Panels Probe Gov’t Use of Facial Recognition Software

https://www.securityweek.com/tech-industry-leaders-endorse-regulating-artificial-intelligence-at-rare-summit-in-washington/




AI can now generate CD-quality music from text, and it’s only getting better

A 3D illustration of a toy robot singing.

Imagine typing “dramatic intro music” and hearing a soaring symphony or writing “creepy footsteps” and getting high-quality sound effects. That’s the promise of Stable Audio, a text-to-audio AI model announced Wednesday by Stability AI that can synthesize stereo 44.1 kHz music or sounds from written descriptions. Before long, similar technology may challenge musicians for their jobs.

If you’ll recall, Stability AI is the company that helped fund the creation of Stable Diffusion, a latent diffusion image synthesis model released in August 2022. Not content to limit itself to generating images, the company branched out into audio by backing Harmonai, an AI lab that launched music generator Dance Diffusion in September.

Now Stability and Harmonai want to break into commercial AI audio production with Stable Audio. Judging by production samples, it seems like a significant audio quality upgrade from previous AI audio generators we’ve seen.

On its promotional page, Stability provides examples of the AI model in action with prompts like “epic trailer music intense tribal percussion and brass” and “lofi hip hop beat melodic chillhop 85 bpm.” It also offers samples of sound effects generated using Stable Audio, such as an airline pilot speaking over an intercom and people talking in a busy restaurant.

To train its model, Stability partnered with stock music provider AudioSparx and licensed a data set “consisting of over 800,000 audio files containing music, sound effects, and single-instrument stems, as well as corresponding text metadata.” After feeding 19,500 hours of audio into the model, Stable Audio knows how to imitate certain sounds it has heard on command because the sounds have been associated with text descriptions of them within its neural network.

A block diagram of the Stable Audio architecture provided by Stability AI.
Enlarge / A block diagram of the Stable Audio architecture provided by Stability AI.
Stablility AI

Stable Audio contains several parts that work together to create customized audio quickly. One part shrinks the audio file down in a way that keeps its important features while removing unnecessary noise. This makes the system both faster to teach and quicker at creating new audio. Another part uses text (metadata descriptions of the music and sounds) to help guide what kind of audio is generated.

To speed things up, the Stable Audio architecture operates on a heavily simplified, compressed audio representation to reduce inference time (the amount of time it takes for a machine learning model to generate an output once it has been given an input). According to Stability AI, Stable Audio can render 95 seconds of 16-bit stereo audio at a 44.1 kHz sample rate (often called “CD quality“) in less than one second on an Nvidia A100 GPU. The A100 is a beefy data center GPU designed for AI use, and it’s far more capable than a typical desktop gaming GPU.

As mentioned, Stable Audio isn’t the first music generator based on latent diffusion techniques. Last December, we covered Riffusion, a hobbyist take on an audio version of Stable Diffusion, though its resulting generations were far from Stable Audio’s samples in quality. In January, Google released MusicLM, an AI music generator for 24 kHz audio, and Meta launched a suite of open source audio tools (including a text-to-music generator) called AudioCraft in August. Now, with 44.1 kHz stereo audio, Stable Diffusion is upping the ante.

Stability says Stable Audio will be available in a free tier and a $12 monthly Pro plan. With the free option, users can generate up to 20 tracks per month, each with a maximum length of 20 seconds. The Pro plan expands these limits, allowing for 500 track generations per month and track lengths of up to 90 seconds. Future Stability releases are expected to include open source models based on the Stable Audio architecture, as well as training code for those interested in developing audio generation models.

As it stands, it’s looking like we might be on the edge of production-quality AI-generated music with Stable Audio, considering its audio fidelity. Will musicians be happy if they get replaced by AI models? Likely not, if history has shown us anything from AI protests in the visual arts field. For now, a human can easily outclass anything AI can generate, but that may not be the case for long. Either way, AI-generated audio may become another tool in a professional’s audio production toolbox.

https://arstechnica.com/?p=1968056




Coca-Cola embraces controversial AI image generator with new “Y3000” flavor

A promotional image for Coca-Cola Y3000

Coca-Cola has taken a fizzy leap into the future of AI hype with the release of Coca‑Cola Y3000 Zero Sugar, a “limited-edition” beverage reportedly co-created with artificial intelligence. Its futuristic name evokes flavor in the year 3000 (still 977 years away), but its marketing relies on AI-generated imagery from 2023—courtesy of the controversial image synthesis model Stable Diffusion.

Stable Diffusion, a technology that is mentioned by name when launching the “Coca-Cola Y3000 AI Cam” mobile app, gained its ability to generate images by scraping hundreds of millions of copyrighted works found on the Internet without copyright-holder permission and is currently the subject of litigation related to copyright infringement.

But there is no hint of that controversy in Coca-Cola’s marketing materials, which lean heavily into today’s buzzy, AI-centered tech zeitgeist.

“We hope that Coca‑Cola will still be as relevant and refreshing in the year 3000 as it is today, so we challenged ourselves to explore the concept of what a Coke from the future might taste like—and what kind of experiences would a Coke from the future unlock?” Oana Vlad, Coke’s senior director of global strategy, said in a press release.

An AI-augmented QR code provided by Coca-Cola.
Enlarge / An AI-augmented QR code provided by Coca-Cola.

To explore that hypothetical future, Coca-Cola developed a mystery flavor and wrapped it in an AI-themed marketing package and mobile web-based app that draws upon both human designers and the aforementioned Stable Diffusion. There’s no direct reference to AI being used to develop the flavor beyond this one statement: “Fans’ perspectives from around the world, combined with insights gathered from artificial intelligence, helped inspire Coca‑Cola to create the unique taste of Y3000.”

The lack of an AI mixologist at the helm is probably a good thing because AI has been known to concoct questionable recipes. Nobody needs toxic ingredients accidentally confabulated into their soda recipe. (Would Coca-Cola feed its vaulted top-secret recipe through a commercial cloud AI model, such as ChatGPT, anyway?)

While there may be no AI flavors involved, the company talks about using AI to develop Y3000’s graphical branding: “Coca‑Cola Y3000 Zero Sugar sports an equally futuristic—and optimistic—visual identity. Co-created with artificial intelligence, the design showcases liquid in a morphing, evolving state, communicated through form and color changes that emphasize a positive future.”

The graphic design is not the only place Stable Diffusion plays a part in the Y3000 marketing package. On its website, Coca-Cola provides an AI-augmented QR code (likely generated with a recently discovered Stable Diffusion technique). When scanned on a mobile phone, the code takes thirsty users to the “Coca‑Cola Creations Hub,” where they can take photos with their device camera and filter them through Stable Diffusion to have a vaporwave-style futuristic look.

As it stands today, the term “artificial intelligence” is a technical term of art, yet it’s also arguably sometimes meaningless due to misunderstandings and misuse. Now that it has become marketing jargon for a multinational food and beverage company, the trend toward watering down AI’s meaning seems likely to continue.

Coca-Cola mentions that the Y3000 AI Cam is powered by Stable Diffusion on its website.
Enlarge / Coca-Cola mentions that the Y3000 AI Cam is powered by Stable Diffusion on its website.
Benj Edwards

Coca-Cola says that the zero-sugar version of the new AI-augmented soda will be available for a limited time in “select markets” including the United States, Canada, China, Europe, and Africa. Thirsty futuristic folks in the US, Canada, and Mexico will also be able to buy an “original taste version” of Coca‑Cola Y3000 soon.

https://arstechnica.com/?p=1967928




AI, la mappa dei nuovi lavori. Uno sguardo globale

Il mercato globale dei lavori basati sull’AI è in fase di boom, con 97 milioni di nuovi profili entro il 2025. Ma cresce nel contempo la preoccupazione per il cambio di mansioni per la forza lavoro. L’intero panorama del mondo del lavoro è in fase di mutamento con la necessità di nuove competenze per adattarsi alla domanda di nuovi profili. Fra il 2022 e il 2023 la domanda di posizioni basate sull’AI è cresciuta del 450%, imponendo un’accelerazione nel processo di trasformazione delle skill lavorative. Se da un lato resta la paura di una sostituzione tout court dell’AI rispetto al lavoro umano, nuove opportunità stanno emergendo in questo contesto in rapido movimento. A seguire uno sguardo sui paesi dove la domanda di AI è più consistente.

Elenchi di lavoro AI in tutto il mondo

Le opportunità di lavoro legate all’intelligenza artificiale sono aumentate negli ultimi anni, con proiezioni che indicano che entro il 2025, 85 milioni di posti di lavoro potrebbero essere spostati, ma potrebbero emergere ben 97 milioni di nuovi ruoli per sostituirli. Uno sguardo al mercato del lavoro dell’intelligenza artificiale a partire dall’agosto 2023 rivela una dinamica.

Image courtesy of Executive Placements.

1. Gli Stati Uniti sono in cima alle classifiche – Gli Stati Uniti sono in testa alla classifica con ben 75.685 annunci di lavoro nel campo dell’intelligenza artificiale, riaffermando la loro posizione di potenza globale nel campo dell’intelligenza artificiale.

2. Domanda in crescita in India – L’India è al secondo posto, offrendo 11.855 posti di lavoro nel settore dell’intelligenza artificiale, a testimonianza del boom del suo settore tecnologico.

3. Il Tech Surge del Giappone _ Il Giappone segue da vicino, con 8.157 offerte di lavoro nel settore dell’intelligenza artificiale, evidenziando il suo impegno per l’innovazione dell’intelligenza artificiale.

4. Il Regno Unito si sta ritagliando una nicchia _ Il Regno Unito presenta 5.837 annunci di lavoro legati all’intelligenza artificiale, sottolineando il suo ruolo nel panorama globale dell’intelligenza artificiale.

5. La Francia in gara – La Francia, con 3.984 offerte di lavoro legate all’IA, consolida la sua presenza come attore dell’IA.

6. Il contributo del Canada – Il Canada si assicura la sesta posizione, offrendo 2.566 posti di lavoro nel settore dell’intelligenza artificiale, sottolineando il suo impegno per lo sviluppo dell’intelligenza artificiale.

7. L’ascesa dell’Australia – L’Australia elenca 937 posti di lavoro nel campo dell’intelligenza artificiale, dimostrando la crescita del suo settore dell’intelligenza artificiale.

8. L’America Latina si unisce – Il Messico fa sentire la sua presenza rispettivamente con 909 annunci di lavoro AI.

9. Israele – Israele vanta 485 opportunità di lavoro nel settore dell’intelligenza artificiale, contribuendo alla diversità e alla crescita del mercato globale del lavoro nel settore dell’intelligenza artificiale.

10. L’ingresso del Sudafrica – Il Sudafrica richiede 429 posti di lavoro legati all’intelligenza artificiale, indicando una scena dell’intelligenza artificiale nascente ma promettente.

Questo mercato del lavoro dinamico contrasta con la situazione del giugno 2023, quando gli Stati Uniti avevano richiesto l’incredibile cifra di 173.000 esperti legati all’intelligenza artificiale. L’India ha cercato 259.000 professionisti dell’IA, sottolineando la rapida crescita del settore.

L’impatto globale dell’IA sull’occupazione

Alcuni paesi avvertono la pressione del potenziale dell’intelligenza artificiale di sostituire i posti di lavoro. L’Ungheria prevede che il 36,4% della sua forza lavoro sarà sostituita dall’intelligenza artificiale, seguita da vicino da Repubblica Ceca (35%) e Polonia (33,5%). L’impatto globale è significativo, con Portogallo (30,1%), Estonia (29,8%), Germania (28,7%), Francia (27,4%), Irlanda (22,5%), Stati Uniti (21,2%) e Regno Unito (19% ) prevedendo inoltre un sostanziale spostamento di posti di lavoro.

Disparità regionali negli Stati Uniti

Negli Stati Uniti, circa l’1,5% delle offerte di lavoro in qualsiasi stato sono legate all’intelligenza artificiale. Ma le disparità regionali sono evidenti. Nel 2022, la California e il Texas vantavano la più alta domanda di lavoratori dell’intelligenza artificiale, seguiti da New York, Virginia, Ohio, Florida, Georgia e Illinois. In netto contrasto, il Montana è rimasto indietro con solo lo 0,72% di offerte di lavoro basate sull’intelligenza artificiale nel paese.

Image courtesy of Executive Placements.

Il mercato del lavoro dell’IA è diversificato, con diversi ruoli che conosceranno una crescita notevole tra il 2021 e il 2031:

1. Data Scientist: in testa alla classifica con una notevole crescita del 36%.

2. Sviluppatori di software: seguono da vicino, con un tasso di crescita del 25%.

3. Scienziati di ricerca informatica e informatica: un tasso di crescita del 21%.

4. Informatica generale e tecnologia dell’informazione – Con un tasso di crescita del 15%.

La richiesta di competenze specifiche in materia di intelligenza artificiale è in aumento, con competenze in elaborazione del linguaggio naturale (NLP), sviluppo di algoritmi, PyTorch, apprendimento automatico e MS Azure ML particolarmente ricercate nel mercato del lavoro. Queste competenze specialistiche sono fondamentali per sfruttare la potenza dell’intelligenza artificiale per risolvere problemi complessi e promuovere l’innovazione in vari settori. Mentre l’intelligenza artificiale continua ad evolversi, i professionisti competenti in queste aree sono pronti a svolgere un ruolo fondamentale nel plasmare il futuro della tecnologia.

Casi d’uso emergenti dell’IA

L’intelligenza artificiale sta orchestrando un’ondata di trasformazione in tutti i settori, non solo modificando l’occupazione ma anche alimentando l’innovazione. I casi d’uso dell’IA adottati più rapidamente comprendono l’ottimizzazione delle operazioni di servizio, in testa al 24%, seguita da vicino dalla creazione di nuovi prodotti basati sull’intelligenza artificiale al 20%. Inoltre, l’intelligenza artificiale sta contribuendo attivamente alla segmentazione dei clienti e all’analisi dei servizi, entrambe al 19%, migliorando allo stesso tempo i prodotti attraverso le funzionalità dell’intelligenza artificiale, sottolineando ulteriormente il suo impatto multiforme sulle aziende moderne.

Settori che attraggono investimenti nell’IA

Media, piattaforme social e marketing sono emersi in prima linea, attirando l’incredibile cifra di 46 miliardi di dollari di finanziamenti. Questa infusione di capitale sottolinea l’impegno del settore nel rivoluzionare la pubblicità, la creazione di contenuti e il coinvolgimento dei consumatori attraverso strategie basate sull’intelligenza artificiale.

I servizi finanziari e assicurativi hanno seguito da vicino, assicurando un sostanziale investimento di 32 miliardi di dollari. Ciò riflette il vivo interesse del settore nello sfruttare il potenziale dell’intelligenza artificiale per migliorare la valutazione del rischio, l’individuazione delle frodi e il servizio clienti, rimodellando così il panorama finanziario.

Nel settore delle infrastrutture IT e dell’hosting, sono stati investiti ingenti 13 miliardi di dollari nell’intelligenza artificiale, segnando un chiaro impegno verso l’ottimizzazione dei data center, della sicurezza informatica e dei servizi cloud utilizzando l’intelligenza artificiale.

Il settore della mobilità e dei veicoli autonomi ha ricevuto 12 miliardi di dollari in investimenti nell’intelligenza artificiale, confermando l’attenzione del settore sui trasporti intelligenti e sulla tecnologia di guida autonoma per migliorare la sicurezza e l’efficienza.

Sanità, farmaci e biotecnologia hanno raccolto 8 miliardi di dollari, dimostrando il ruolo crescente dell’intelligenza artificiale nel rivoluzionare la scoperta di farmaci, la cura dei pazienti e la ricerca medica, con il potenziale di far avanzare la diagnostica e i trattamenti.

I processi aziendali e i servizi di supporto hanno visto un notevole investimento di 5 miliardi di dollari nell’intelligenza artificiale. Questo finanziamento riflette l’impegno del settore a sfruttare l’intelligenza artificiale per semplificare le operazioni, migliorare l’assistenza ai clienti e aumentare l’efficienza complessiva in vari processi aziendali. Questi investimenti significativi evidenziano collettivamente l’impatto trasformativo dell’intelligenza artificiale in diversi settori del mercato globale.

Sei pronto per il futuro del lavoro?

Poiché l’adozione dell’IA per le organizzazioni globali raggiungerà il 55% nel 2023, è evidente che i posti di lavoro legati all’intelligenza artificiale sono in aumento. Il mercato del lavoro è dinamico, con variazioni regionali e ruoli diversi. Che tu sia un data scientist, uno sviluppatore di software o un appassionato di intelligenza artificiale, le opportunità abbondano in questo fiorente settore. L’aumento del 450% delle offerte di lavoro legate all’intelligenza artificiale tra il 2022 e il 2023 testimonia la crescente importanza dell’intelligenza artificiale nel mercato del lavoro di oggi. La domanda rimane: sei pronto a unirti alla rivoluzione dell’intelligenza artificiale?

https://www.key4biz.it/ai-la-mappa-dei-nuovi-lavori-uno-sguardo-globale/458846/




Microsoft offers legal protection for AI copyright infringement challenges

A man in an armor helmet sitting at a desk with a protective glowing field around him.

On Thursday, Microsoft announced that it will provide legal protection for customers who are sued for copyright infringement over content generated by the company’s AI systems. This new policy, called the Copilot Copyright Commitment, is an expansion of Microsoft’s existing intellectual property indemnification coverage, Reuters reports.

Microsoft’s announcement comes as generative AI tools like ChatGPT have raised concerns about reproducing copyrighted material without proper attribution. Microsoft has heavily invested in AI through products like GitHub Copilot and Bing Chat that can generate original code, text, and images on demand. Its AI models have gained these capabilities by scraping publicly available data off of the Internet without seeking express permission from copyright holders.

By offering legal protection, Microsoft aims to give customers confidence in deploying its AI systems without worrying about potential copyright issues. The policy covers damages and legal fees, providing customers with an added layer of protection as generative AI sees rapid adoption across the tech industry.

“As customers ask whether they can use Microsoft’s Copilot services and the output they generate without worrying about copyright claims, we are providing a straightforward answer: yes, you can, and if you are challenged on copyright grounds, we will assume responsibility for the potential legal risks involved,” writes Microsoft.

Under the new commitment, Microsoft will pay any legal damages for customers using Copilot, Bing Chat, and other AI services as long as they use built-in guardrails.

“Specifically, if a third party sues a commercial customer for copyright infringement for using Microsoft’s Copilots or the output they generate, we will defend the customer and pay the amount of any adverse judgments or settlements that result from the lawsuit, as long as the customer used the guardrails and content filters we have built into our products,” writes Microsoft.

With the rise of generative AI, the tech industry has been grappling with questions about properly crediting or licensing copyrighted source material used in training AI models. Legal experts say these thorny copyright issues will likely be decided through future legislation and court cases, some of which are already underway.

In fact, Microsoft has already attracted litigation over Copilot technology. Last November, the Joseph Saveri Law Firm filed a class-action lawsuit against Microsoft and OpenAI over GitHub Copilot’s alleged copyright violations that arose from scraping publicly available code repositories. Currently, the status of that lawsuit is unknown, and Ars Technica could not confirm if the case is still active using public records.

https://arstechnica.com/?p=1966332




The AI-assistant wars heat up with Claude Pro, a new ChatGPT Plus rival

The Anthropic Claude logo on a purple background.
Enlarge / The Anthropic Claude logo.
Anthropic / Benj Edwards

On Thursday, AI-maker and OpenAI competitor Anthropic launched Claude Pro, a subscription-based version of its Claude.ai web-based AI assistant, which functions similarly to ChatGPT. It’s available for $20/month in the US or 18 pounds/month in the UK, and it promises five-times-higher usage limits, priority access to Claude during high-traffic periods, and early access to new features as they emerge.

Like ChatGPT, Claude Pro can compose text, summarize, do analysis, solve logic puzzles, and more.

Claude.ai is what Anthropic offers as its conversational interface for its Claude 2 AI language model, similar to how ChatGPT provides an application wrapper for the underlying models GPT-3.5 and GPT-4. In February, OpenAI chose a subscription route for ChatGPT Plus, which for $20 a month also gives early access to new features, but it also unlocks access to GPT-4, which is OpenAI’s most powerful language model.

What does Claude have that ChatGPT doesn’t? One big difference is a 100,000-token context window, which means it can process about 75,000 words at once. Tokens are fragments of words used while processing text. That means Claude can analyze longer documents or hold longer conversations without losing its memory of the subject at hand. ChatGPT can only process about 8,000 tokens in GPT-4 mode.

“The big thing for me is the 100,000-token limit,” AI researcher Simon Willison told Ars Technica. “That’s huge! Opens up entirely new possibilities, and I use it several times a week just for that feature.” Willison regularly writes about using Claude on his blog, and he often uses it to clean up transcripts of his in-person talks. Although he also cautions about “hallucinations,” where Claude sometimes makes things up.

“I’ve definitely seen more hallucinations from Claude as well” compared to GPT-4, says Willison, “which makes me nervous using it for longer tasks because there are so many more opportunities for it to slip in something hallucinated without me noticing.”

A screenshot of the public web interface for Claude 2, a large language model from Anthropic.
Enlarge / A screenshot of the public web interface for Claude 2, a large language model from Anthropic.
Ars Technica

Willison has also run into problems with Claude’s morality filter, which has caused him trouble by accident: “I tried to use it against a transcription of a podcast episode, and it processed most of the text before—right in front of my eyes—it deleted everything it had done! I eventually figured out that we had started talking about bomb threats against data centers towards the end of the episode, and Claude effectively got triggered by that and deleted the entire transcript.”

What does “5x more usage” mean?

Anthropic’s primary selling point for the Claude Pro subscription is “5x more usage,” but the company doesn’t clearly communicate what Claude’s free-tier usage limits actually are. Dropping clues like cryptic breadcrumbs, the company has written a support document about the topic that says, “If your conversations are relatively short (approximately 200 English sentences, assuming your sentences are around 15–20 words), you can expect to send at least 100 messages every 8 hours, often more depending on Claude’s current capacity. Over two-thirds of all conversations on claude.ai (as of September 2023) have been within this length.”

In another somewhat cryptic statement, Anthropic writes, “If you upload a copy of The Great Gatsby, you may only be able to send 20 messages in that conversation within 8 hours.” We’re not attempting the math, but if you know the precise word count of F. Scott Fitzgerald’s classic, it may be possible to glean Claude’s actual limits. We reached out to Anthropic for clarification yesterday and have not received a response before publication.

Either way, Anthropic makes it clear that for an AI model with a 100,000-token context limit, usage limits are necessary because the computation involved is expensive. “A model as capable as Claude 2 takes a lot of powerful computers to run, especially when responding to large attachments and long conversations,” Anthropic writes in the support document. “We set these limits to ensure Claude can be made available to many people to try for free, while allowing power users to integrate Claude into their daily workflows.”

Also, on Friday, Anthropic launched Claude Instant 1.2, which is a less expensive and faster version of Claude that is available through an API. However, it’s less capable than Claude 2 in logical tasks.

While it’s clear that large language models like Claude can do interesting things, it seems their drawbacks and tendency toward confabulation may hold them back from wider use due to reliability concerns. Still, Willison is a fan of Claude 2 in its online form: “I’m excited to see it continue to improve. The 100,000-token thing is a massive win, plus the fact you can upload PDFs to it is really convenient.”

https://arstechnica.com/?p=1966121




OpenAI confirms that AI writing detectors don’t work

A photo of a teacher covering his eyes.

Last week, OpenAI published tips for educators in a promotional blog post that shows how some teachers are using ChatGPT as an educational aid, along with suggested prompts to get started. In a related FAQ, they also officially admit what we already know: AI writing detectors don’t work, despite frequently being used to punish students with false positives.

In a section of the FAQ titled “Do AI detectors work?”, OpenAI writes, “In short, no. While some (including OpenAI) have released tools that purport to detect AI-generated content, none of these have proven to reliably distinguish between AI-generated and human-generated content.”

In July, we covered in depth why AI writing detectors such as GPTZero don’t work, with experts calling them “mostly snake oil.” These detectors often yield false positives due to relying on unproven detection metrics. Ultimately, there is nothing special about AI-written text that always distinguishes it from human-written, and detectors can be defeated by rephrasing. That same month, OpenAI discontinued its AI Classifier, which was an experimental tool designed to detect AI-written text. It had an abysmal 26 percent accuracy rate.

OpenAI’s new FAQ also addresses another big misconception, which is that ChatGPT itself can know whether text is AI-written or not. OpenAI writes, “Additionally, ChatGPT has no ‘knowledge’ of what content could be AI-generated. It will sometimes make up responses to questions like ‘did you write this [essay]?’ or ‘could this have been written by AI?’ These responses are random and have no basis in fact.”

Along those lines, OpenAI also addresses its AI models’ propensity to confabulate false information, which we have also covered in detail at Ars. “Sometimes, ChatGPT sounds convincing, but it might give you incorrect or misleading information (often called a ‘hallucination’ in the literature),” the company writes. “It can even make up things like quotes or citations, so don’t use it as your only source for research.”

(In May, a lawyer got in trouble for doing just that—citing six non-existent cases that he pulled from ChatGPT.)

Even though automated AI detectors do not work, that doesn’t mean a human can never detect AI writing. For example, a teacher familiar with a student’s typical writing style can tell when their style or capability suddenly changes. Also, some sloppy attempts to pass off AI-generated work as human-written can leave tell-tale signs, such as the phrase “as an AI language model,” which means someone copied and pasted ChatGPT output without being careful. And recently, an article in the scientific journal Nature showed how humans noticed the phrase “Regenerate response” in a scientific paper, which is the label of a button in ChatGPT.

As the technology stands today, it’s safest to avoid automated AI detection tools completely. “As of now, AI writing is undetectable and likely to remain so,” frequent AI analyst and Wharton professor Ethan Mollick told Ars in July. “AI detectors have high false positive rates, and they should not be used as a result.”

https://arstechnica.com/?p=1966483




Intelligenza artificiale, la Spagna prima in Europa a istituire una agenzia ad hoc. Lo statuto

Avrà sede a Lo Coruna, in Galizia, dopo una gara con altre 16 città in lizza su tutto il territorio, la nuova agenzia spagnola dedicata interamente alla supervisione dell’Intelligenza Artificiale. Un modo intelligente per valorizzare una regione, la Galizia, che diventa così la culla dell’IA in Spagna, nuovo centro di gravità della tecnologia europea.

Il decreto regio per la nuova agenzia è stato siglato il 22 agosto.

L’obiettivo della Agenzia per la supervisione dell’Intelligenza Artificiale (AESIA) è quello di garantire in linea con la strategia digitale nazionale che lo sviluppo dell’AI nel paese sia “inclusivo, sostenibile e incentrato sui cittadini”.

L’agenzia sarà formata in collaborazione fra il ministero spagnolo delle finanze e del servizio civile e quello degli affari economici e della trasformazione digitale.

L’agenzia rientra nel più ampio piano per una strategia nazionale per l’intelligenza artificiale. Secondo una nota, la Spagna diventa così il primo paese europeo a creare una agenzia ad hoc per l’AI.

Scarica qui lo statuto dell’Agenzia spagnola per l’Intelligenza Artificiale in PDF

Funzioni dell’Agenzia

Fra le sue funzioni, la supervisione dell’uso di questa tecnologia nei suoi diversi campi, nonché il controllo della commercializzazione dei sistemi che includono l’intelligenza artificiale (AI) e, in particolare, quelli che possono comportare rischi significativi per la salute, la sicurezza e diritti individuali.

L’agenzia intende inoltre fungere da trampolino di lancio per l’introduzione e il buon utilizzo verificabile e aperto dell’IA; la definizione di meccanismi consultivi e assistenziali; la collaborazione e il coordinamento con altre autorità di vigilanza sull’IA (nazionali e sovranazionali) e la promozione di veri e propri ambienti di test per i sistemi di IA per rafforzare la tutela degli utenti.

AI Act approvato a giugno

A giugno l’Unione Europea ha approvato l’AI Act, il quadro normativo di base per la governance e la gestione dell’AI. Una volta recepito, il nuovo quadro di regole imporrà una serie di restrizioni su diversi servizi basati sull’AI, proibendo diversi prodotti.

Tra le tecnologie che potrebbero essere completamente bandite ci sono la sorveglianza biometrica, i sistemi di punteggio sociale, gli algoritmi di polizia predittiva, i software di riconoscimento delle emozioni e i sistemi di riconoscimento facciale non mirati.

Tuttavia, i modelli di intelligenza artificiale generativa come ChatGPT di OpenAI e Bard di Google potrebbero funzionare, a condizione che i loro risultati siano chiaramente contrassegnati come generati dall’intelligenza artificiale. L’AI Act obbliga tutti i paesi dell’Unione a istituire agenzie nazionali che supervisionano l’intelligenza artificiale. A questo proposito, l’intervista a Marco Scialdone, Head Litigation and Academic Outreach di Euroconsumers.

Scarica qui lo statuto dell’Agenzia spagnola per l’Intelligenza Artificiale in PDF

https://www.key4biz.it/intelligenza-artificiale-la-spagna-prima-in-europa-a-istituire-una-agenzia-ad-hoc-lo-statuto/458741/