Kai kurie viršininkai nelinkę plėsti generatyvaus dirbtinio inteleko naudojimą, tuo tarpu jų darbuotojai yra pilnai įsitraukę į asmeninį naudojimą

 „Investuotojus, susirūpinusius, kad Amerikos technologijų gigantai beatodairiškai stato ant generatyvaus dirbtinio intelekto (AI), naujausi didžiųjų technologijų ketvirčio rezultatai suteikė šiek tiek užtikrintumo. „Amazon“, „Microsoft“ ir „Google“ debesijos paslaugų paklausos augimas iš įmonių buvo labai didelis. Andy Jassy, ​​„Amazon“ vadovas, teigė, kad AI pajamos iš „Amazon Web Services“ (AWS) auga triženkliais rodikliais – tris kartus greičiau, nei pati AWS išaugo pirmaisiais metais po to, kai 2006 m. pradėjo kurti debesų kompiuteriją.

 

 Vis dėlto pasinerkite giliau ir situacija yra labiau niuansuota. Atrodo, kad „generatyvus AI“ yra viena iš tų naujovių, kaip, pvz., el. paštas ar išmanieji telefonai, kuriais anksti pradeda naudoti atskiri asmenys.

 

 Įmonės yra daug atsargesnės.

 

 Per dvejus metus nuo tada, kai „OpenAi“ pristatė „ChatGPT“, generatyvusis AI buvo naudojamas greičiau, nei kompiuteriai ar internetas. Remiantis Aleksandro Bicko iš Sent Luiso federalinio rezervo banko ir bendraautorių tyrimu, visiškai 39% amerikiečių sako, kad jie jį naudoja; 28 % teigia, kad naudojasi darbe, o 11 % – kad tai daro kasdien.

 

 Tačiau daugelis iš jų, atrodo, yra slapti kiborgai, naudojasi technologijomis darbe, net kai jų darbdaviai blaškosi. Remiantis JAV gyventojų surašymo biuro apklausa, tik 5% Amerikos įmonių teigia, kad naudoja technologiją prekėms ar paslaugoms gaminti. Atrodo, kad daugelis kompanijų kenčia nuo ūmaus eksperimentito atvejo, o ne iki galo įgyvendindamos technologiją, bandančios įgyvendinti tik bandomuosius projektus. Neseniai 14 šalių profesionalių paslaugų įmonės Deloitte atliktoje apklausoje tik 8 % įmonių vadovų teigė, kad jų įmonės panaudojo daugiau, nei pusę generatyvaus AI eksperimentų.

 

 Dėl to pajamos iš AI paslaugų pardavimo įmonėms išlieka ribotos. Nors ponas Jassy sakė, kad AWS dabar iš AI generuoja „kelis milijardus“ dolerių pajamų, tai yra tik nedidelė dalis iš 110 mlrd. dolerių visų jos pajamų. Konsultacijų milžinė „Accenture“, neseniai paskelbusi, kad apmokys 30 000 darbuotojų, kad padėtų įmonėms pritaikyti generatyvųjį dirbtinį intelektą, rugsėjį pranešė, kad per pastaruosius 12 mėnesių ji užsakė su šia technologija susijusių darbų už 3 mlrd. dolerių, ty dešimt kartų daugiau, nei prieš metus. Tačiau, palyginti su bendrais bendrovės pardavimais, daugiau nei 81 mlrd. dolerių, tai taip pat yra mažas alus.

 

 Kodėl daugelis viršininkų nesiryžta priimti generatyvaus AI? Viena iš priežasčių yra ta, kad jie nerimauja dėl neigiamų pusių. Klausykitės technologijų gigantų ir jie jums pasakys, kaip liepą pasakė Sundaras Pichai, „Alphabet“ vadovas, kad „per mažai investavimo rizika yra žymiai didesnė, nei per didelio investavimo rizika“. Tikimasi, kad „Alphabet“, „Amazon“, „Microsoft“ ir „Meta“ šiais metais su AI susijusioms kapitalo išlaidoms skirs mažiausiai 200 mlrd. dolerių. 

 

Kitų pramonės šakų viršininkai yra atsargesni. Neseniai vykusioje diskusijoje už uždarų durų didelės Amerikos verslo grupės vadovas kalbėjo apie dviejų tipų vadovų baimes dėl generatyvaus AI. Vienas būti paliktu nuošalyje, jei jį priimi per lėtai. Kitas būti sugėdintu, jei per greitai judi ir pakenki savo įmonės reputacijai.

 

 Teisinė ir reguliavimo rizika yra didelė. Ieškiniai, susiję su privatumu, šališkumu ir autorių teisių pažeidimais, patenka į teismus. Rugpjūčio mėnesį įsigaliojo Europos Sąjungos AI įstatymas. Šiais metais dirbtinio intelekto įstatymai buvo pateikti mažiausiai 40 Amerikos valstijų. Stipriai reguliuojamų pramonės šakų, tokių, kaip sveikatos priežiūra ir finansai, vadovai yra ypač atsargūs.

 

 Nors jie mato generatyvaus dirbtinio intelekto potencialą pakeisti jų verslą, tarkim, pagreitinant vaistų atradimą ar sukčiavimo aptikimą, jie puikiai žino apie grėsmes privatumui ir saugumui, jei būtų pažeisti jų klientų medicininiai ar finansiniai duomenys.

 

 Kita problema yra ta, kad generatyvaus AI pritaikymo nauda gali būti neaiški. Prieiga prie didelių kalbos modelių (LLM) yra brangi, nesvarbu, ar per įmonės serverius (saugesnis kelias), ar per debesies paslaugų teikėjus (paprastesnis kelias).

 

 Viso masto generatyvaus AI įdiegimas gali padidinti pajamas ir sumažinti išlaidas, tačiau atsipirkimas nėra iš karto, todėl kyla susirūpinimas dėl investicijų grąžos. Neseniai atliktoje tyrime Deloitte nustatė, kad vyresniųjų vadovų, kurie „didelį“ arba „labai didelį“ susidomėjimą generuojančiu dirbtiniu intelektu, dalis sumažėjo iki 63 %, palyginti su 74 % pirmąjį metų ketvirtį, o tai rodo, kad naujų technologijų spindesys“ gali susilpnėti. Vienas vadovas apibendrina skepticizmą, papasakodamas istoriją apie vyriausiąjį informacijos pareigūną, kurio viršininkas liepė liautis žadėjus 20 % produktyvumo didinimą, nebent iš pradžių būtų pasirengęs penktadaliu sumažinti savo skyriaus darbuotojų skaičių.

 

 Net tada, kai įmonės nori sustiprinti generatyvinio AI naudojimą, joms tai gali pasirodyti sudėtinga. „Accenture“ dirbtinio intelekto vadovas Lan Guanas, kad išnaudotų visą technologijos naudą, pirmiausia turi sutvarkyti savo duomenis, sistemas ir darbo jėgą. Ji skaičiuoja, kad įmonių pasirengimas generatyviniam AI yra daug mažesnis, nei ankstesnių technologijų bangų, tokių, kaip internetas ar debesų kompiuterija, atveju.

 

 Viena iš problemų yra netvarkingi duomenys, išsibarstę įvairiais formatais įvairiuose skyriuose ir programinės įrangos sistemose. Ponia Guan pateikia pavyzdį apie telekomunikacijų įmonę, kuri norėjo išmokyti skambučių centro AI padėjėją, tiekdama jai PDF failus, vadovus, skambučių žurnalus ir kt. Botas išsiaiškino, kad vietoj vienos standartinės veiklos procedūros, kurią ji vadina „vienu tiesos šaltiniu“, įmonė turėjo 37, sukauptus per dešimtmečius. Ji sako, kad nesugebėjimas sutvarkyti duomenų, prieš naudojant juos botui mokyti, padidina haliucinacijų ir klaidų riziką.

 

 Kita problema yra ta, kad IT sistemos dažnai yra girgždančios ir pasenusios. Ši problema vadinama „technine skola“. Dėl to gali būti sunku prijungti LLM, nesukeliant problemų. Pusiau autonominių AI agentų integravimas į, žmonėms sukurtas, sistemas taip pat gali sukurti saugumo spragų.

 

 Tada yra įgūdžių problema. Daugelis įmonių vis dar stengiasi rasti pakankamai dirbtinio intelekto specialistų. Tyrimų įmonės „Lightcast“ duomenimis, su dirbtiniu intelektu susijusių, darbo skelbimų skaičius Amerikoje šiais metais išaugo 122%, palyginti su 18% padidėjimu 2023 m. „Lightcast“ ekonomistė Elizabeth Crofoot teigia, kad šį padidėjimą daugiausia paaiškina generatyvus AI, darbo aprašymuose minimas „ChatGPT“, sparti inžinerija ir spartus kalbų modeliavimas.

 

 Įmonės taip pat nori kitų pareigų darbuotojų, kurie žinotų, kaip naudoti generatyvųjį AI. Pardavėjas, turintis dirbtinio intelekto įgūdžių, gali uždirbti 45 000 dolerių per metus daugiau, nei tas, kuriam jų trūksta, sako ponia Crofoot. Todėl nenuostabu, kad net kai kurie viršininkai nesistengia didinti generatyvaus dirbtinio intelekto naudojimą, jų darbuotojai yra visiškai įsitraukę.“ [1]

 

1.  A nasty case of pilotitis. The Economist; London Vol. 453, Iss. 9422,  (Nov 9, 2024): 58, 59.

As some bosses prevaricate about scaling up generative AI, their employees are all in

 

"For investors concerned that America’s tech giants are making recklessly large bets on generative artificial intelligence (AI), big tech’s latest quarterly results have offered some reassurance. The growth in demand from companies for the cloud services of Amazon, Microsoft and Google was red hot. Andy Jassy, boss of Amazon, said that AI revenue for Amazon Web Services (AWS) was growing at triple-digit rates—three times faster than AWS itself grew in the early years after it pioneered cloud computing in 2006.

Dig deeper, though, and the situation is more nuanced. Generative ai appears to be one of those innovations, such as email or smartphones, whose the most eager early adopters are individuals.

 Companies are being far more tentative.

In the two years since OpenAi unveiled ChatGPT, generative AI has had a faster rate of adoption than PCs or the internet. Fully 39% of Americans now say they use it, according to a study by Alexander Bick of the Federal Reserve Bank of St Louis and co-authors; 28% say they use it for work, and 11% that they do so every day.

Many of them, though, seem to be secret cyborgs, using the technology at work even as their employers dawdle. Just 5% of American businesses say they are using the technology to produce goods or services, according to a survey by the us Census Bureau. Many companies seem to be suffering from an acute case of pilotitis, dilly-dallying with pilot projects without fully implementing the technology. In a recent survey conducted across 14 countries by Deloitte, a professional-services firm, only 8% of company leaders said their firms had deployed more than half of their generative-AI experiments.

As a consequence, revenue from selling AI services to companies remains limited. Although Mr Jassy said AWS now generates “multi-billion” dollars of revenue from AI, that is a smidgen of the $110bn of annual revenue for its cloud business as a whole. Accenture, a consulting giant that recently announced it would train 30,000 staff to help companies adopt generative AI, said in September that it had booked $3bn-worth of work related to the technology over the past 12 months, a ten-fold increase year on year. But compared with the company’s total sales of more than $81bn, that too is small beer.

Why are many bosses hesitating to adopt generative AI? One reason is that they worry about the downsides. Listen to the tech giants and they will tell you—as Sundar Pichai, the boss of Alphabet, said in July—that “the risk of under-investing is dramatically greater than the risk of over-investing”. Alphabet, Amazon, Microsoft and Meta are expected to pour at least $200bn into AI-related capital expenditures this year. Bosses in other industries are more circumspect. At a recent closed-door discussion, the head of a big American business group spoke of two types of fears chief executives have about generative AI. One was being left behind if they adopted it too slowly. The other was being embarrassed if they moved too quickly and damaged their firm’s reputation.

Legal and regulatory risks loom large. Lawsuits related to privacy, bias and copyright violations are making their way through the courts. In August the European Union’s AI Act came into force. AI bills have been introduced in at least 40 American states this year. Bosses in heavily regulated industries, such as health care and finance, are especially wary. 

Although they see the potential of generative AI to transform their businesses, say by speeding up drug discovery or fraud detection, they are keenly aware of the threats to privacy and security if their customers’ medical or financial data are compromised.

Another problem is that the benefits of adopting generative AI can be uncertain. Accessing large language models (LLMs) is expensive, whether via a company’s own servers (safer) or via cloud-service providers (simpler).

 Full-scale implementation of generative AI may increase revenues and reduce costs, but the payoff is not immediate, raising concerns about returns on investment. In its recent survey Deloitte found that the share of senior executives with a “high” or “very high” level of interest in generative AI had fallen to 63%, down from 74% in the first quarter of the year, suggesting that the “new-technology shine” may be wearing off. One executive sums up the scepticism by recounting the story of a chief information officer whose boss told him to stop promising 20% productivity improvements unless he was first prepared to cut his own department’s headcount by a fifth.

Even when companies are eager to amp up their use of generative AI, though, they may find it tricky. To reap the full rewards of the technology, businesses have to first get their data, systems and workforce into shape, says Lan Guan, Accenture’s head of AI. She reckons companies’ readiness for generative AI is much lower than it was for previous technology waves such as the internet or cloud computing.

One problem is messy data, scattered in different formats across various departments and software systems. Ms Guan gives the example of a telecoms firm that wanted to train a call-centre AI assistant by feeding it PDFs, manuals, call logs and more. The bot found that instead of one standard operating procedure—what she calls “a single source of truth”—the company had 37, accumulated over decades. A failure to organise data before using it to train a bot increases the risk of hallucinations and mistakes, she says.

Another problem is that IT systems are often creaky and old, a problem known as “technical debt”. That can make it difficult to plug in LLMs without causing trouble. Integrating semi-autonomous AI agents into systems built for humans might also create security vulnerabilities.

Then there is the problem of skills. Many companies are still struggling to get their hands on enough AI specialists. According to Lightcast, a research firm, AI-related job postings in America have surged by 122% so far this year, compared with an 18% rise in 2023. Elizabeth Crofoot, an economist at Lightcast, says that this increase is mostly explained by generative AI, with job descriptions mentioning ChatGPT, prompt engineering and large language modelling on the rise.

Companies also want workers in other roles who know how to use generative AI. A sales rep with AI skills can earn $45,000 a year more than one who lacks them, says Ms Crofoot. No wonder, then, that even as some bosses prevaricate about scaling up generative AI, their employees are all in." [1]

 

1.  A nasty case of pilotitis. The Economist; London Vol. 453, Iss. 9422,  (Nov 9, 2024): 58, 59.

A Landslide Against the Media

 

"The recriminations are flying, as Barack Obama's and Joe Biden's forces go to war over who's more to blame for Democrats' humiliating defeat Tuesday. So long as the left is pointing fingers, let it direct a big, fat digit at the outfit that played the biggest role in losing it this election: the U.S. media.

That isn't the conventional wisdom, which holds that the press's naked shilling for Democratic candidates amounts to an in-kind campaign contribution. And no doubt the media's ceaseless attacks on Donald Trump and Republicans did help round up some undecided voters. Yet the boosterism for Kamala Harris & Co. came at a far bigger cost: A narrative full of fantasy enabled Democrats to live in a world disconnected from the mood and worries of the country.

Among the most damaging of these fantasies was the four-year press assurance that Joe Biden was sharp as a tack. Even video evidence in June of a confused president wandering aimlessly at the Group of Seven was met with claims that the footage was "edited," "lacking context," "misleading." Only when the Trump-Biden debate made Mr. Biden's decline undeniable did the media drop the charade. Then it immediately turned to recast Ms. Harris -- a presidential primary loser turned unpopular vice president -- as a political genius and the obvious savior of the Democratic Party. How'd that work out?

In a world with a competent press, Mr. Biden's failing constitution would have been front-page news in time for Democrats to confront the unpleasant (yet manageable) reality of needed change. A primary would have produced a tested nominee, likely one less encumbered by the Biden record. As Harris adviser (and Obama veteran) David Plouffe complains that Team Biden created a "hole" too "deep" for his sidekick to dig out of, don't forget the industry whose job it is to call out political fiction, but instead wrote the "Joe Is Fine" novel.

Of course Democrats are shocked that they lost. In a world with a functioning press, the politician who tries to make lemonade out of inflation, crime or border chaos, is slapped as out of touch. In Biden-Harris world, the press printed their spin as gospel. Four years of headlines insisted Americans live in one of the "strongest economies" ever. Crime rates were falling. Red-state governors engaged in "stunts" to magnify the migrant problem. The biggest issues facing our country were climate, systemic racism, abortion and transgender rights.

The fantasies were maintained right up to the election. Even as Republicans pointed to surging voter registration, unprecedented early votes and notable demographic shifts, the headlines insisted that Kamala would claim victory on a wave of abortion-and-Liz Cheney-loving suburban women, comedian-condemning Puerto Ricans, and white dudes impressed by Tim Walz's camo hat. No wonder Tuesday was a surprise. The America that voted for Mr. Trump has never even made an appearance in these outlets.

Democrats now face a choice. On one side are party grown-ups who are publicly acknowledging this defeat as a sharp voter rebuke of progressive policies. They are admitting that lawfare was a mistake, that the party is culturally out of touch, that lunatic interest groups are running the asylum. They worry about a growing political realignment that threatens the party's future. That we are hearing these voices is an improvement over the past eight years.

Yet on the other side are the progressive architects of the mess, already rationalizing away the night as a function of racism, sexism and America's supposed love affair with "fascism." They mark the loss down to "tactical" errors -- the failure to court pro-Palestinian voters, a misallocation of door-knockers, poor timing in ad buys. The party just needs better "messaging" of its "historic achievements."

No surprise, the media is already running with this latter narrative, again providing the party a soothing alternative to the blunt reality of its ideological fail. Will Democrats be lulled again? If they really want to reconnect with voters, they will at some point have to break with what is proving to be a debilitating feedback loop.

The media itself was put on sharp notice this cycle, pushed aside by podcasters and influencers whom voters now trust more to provide reality. Nearly 50 million people have listened to Joe Rogan's interview of Mr. Trump, as it provided a more accurate assessment of the GOP nominee's positions and the concerns of the country than "news" articles about the "authoritarian" intent on destroying the climate, abortion rights, democracy -- choose your obsession.

The Founders accorded the press the honor of inclusion in the First Amendment in recognition of the vital role it plays in keeping pols honest. The industry is meant to ride herd on government -- on both sides -- in the interest of the people. That job is essential -- not only for transparency, but to provide self-deluding politicians constant gut checks as to how their policies sit with the nation. When that guardrail falls, the nation suffers, but so too does the party that gets to live the make-believe." [1]

The press cannot ride herd on Biden camp, since the press has a theory, that only Biden camp is scientific, good, saving the humanity. Voters in Pennsylvania, Russians in Crimea, Georgians in Georgia, according to our press, need some education, then they will come to Biden camp side.

1. A Landslide Against the Media. Strassel, Kimberley A.  Wall Street Journal, Eastern edition; New York, N.Y.. 08 Nov 2024: A.13.