Coexisting with AI isβε☆< no longer a future moment¥₹>ε, it is now the present moment.

Recently, the 2024 Worlε"≠d Artificial Intelligence Confδ"erence successfully concluded in Shang®≈εhai. The three-day exhibition w¥→σ©ill cover an exhibition areδ∑γa of over 52000 square meters, attracti ∑₹ng more than 500 domestic ε±and foreign enterprises and over 1300 '™÷global leaders to partici≠<εpate. The number of exh±​δibits will exceed 1500, and it is expδ ected to drive a total investment↓λ  of over 40 billion yuan. <☆•The specifications, scale, and in₹€"fluence will continue to improve.

As a "barometer" of the domestic AI ind ¶∑±ustry, this conference has made many p♦₩'eople, including journalists, f±αeel that with the continuou♠©<s implementation of big m✘π♥ odels at the application level, AI¥™ has become more down-to-earth. The £₽business opportunities and various π∏transaction volumes brought about≈§ by this conference are also excitin ★λ≠g. But the trends and price d'±isputes of nearly a thou∑•sand large models in China, as well as  €↕the various problems that AI mayΩ∏ cause, have also triggered people&σ↔#39;s thinking and vigilance.

Model implementation, practical‍Ω≤™ity first

Technology must be im"±​plemented in the industry in order t>£o generate value, "saΩ∏id Wu Yunsheng, Vice₩€∑ President of Tencent Cloud, Head of ♠ Tencent Cloud Intelligence, a♥πεγnd Head of Youtu Labor₩≤atory, expressing the consen✔σsus of many attendees:₽π®π compared to last year&#✘♣'↓39;s" thousand model wa♣‍↓r ", enterprises canno✔®t only focus on technology,÷>‌ but also need to deeply grasp the impαλlementation of scenarios.

This is also the beautiful v∑​ision of the Tencent Forum t'✘heme of this year'πσ&βs World Artificial Int≥ ®elligence Conference - "Intelligent C®&reation of the Future".

This year, the development of AI ha♣δσs advanced further, with qualitative ♥©leaps in reasoning and analysis≈←$, creative generation, and emo∑ tional intelligence, achieving the coe‌≥×γxistence of "wisdom" and "spiritual en₩&←★ergy". On the one han¶Ωd, whether it is organizing d $'​ocuments, writing reports, being a foreβ×"→ign teacher, or making videos, it is no↔♣ problem and is a smart an ™d easy-to-use AI assistant; ♥ On the other hand, the emerging thin÷&king abilities have earned ≈£AI the nickname of "human σ≥♣​machine external brain"★×σ.

Such AI, with its fast implementaε•tion and wide range, has l←♥®eft audiences who ha &♣ve participated in conferenceλ♥÷≥s year after year in awe: "Last year, m​ ₩§ost of them were sti♠≥ll conceptual, and in th±♥∑≤e blink of an eye, there w↔δere touchable and perceptible a $$pplications everywhere

A year has passed, and this isλ•✔♠ still a stunning trend, "exhibitors sa"♣id one after anotherε →. A few months ago, the governm∏πent work report at the Nation<₩♣↔al Two Sessions proposed to cβ♥arry out the "AI+" action, w♣✘σhich is similar to the "Internet​δ plus", but has a more far-reachi♦¶ng impact.

It is obvious from the exhibitioε≥≠n that the change from≠↓ "+AI" to "AI+" has brought abδ≈←σout subversive changes - from ↑​$Kwai "Kering" that enables live ©↔♠broadcast to Netease Youdao "Zi‌‌♠yue" that educates and enterta↔®λins, AI has been integrated i♣σ∏♣nto production and life, helping to red×<'™uce costs and increase effic ​↓iency. At the same time, from the ‌®₹©Cybertruck (electric pickup tru≥α•βck) running on the ground to the manned©₹σ eVTOL (electric vertical ±π↔takeoff and landing aircraf↕↓↑t) flying in the sky, it demonstrates ✔' the various possibilities of theγσπ↔ "AI+" future.

Life services, industrial automatio★" n, power inspection, transportation÷← and travel... Large scale models are d∏®←σeeply penetrating various fields of ≈↑various industries, and the ×β÷¥archaeological industry, which is kn>®own as a "niche skill", is no exc∏¶eption. The Hehe Informa'✘→©tion booth uses AI to repair damaged  >✘scrolls, with exquisite details thδ÷φat instantly transpo€¶ ₩rt viewers back to Dunhuan×β₽g; Tencent Forum announced that i¥♠t will officially open source the ±✘world's largest multimodal data♣±₩δset of Oracle bones, allowing rubbin₽φδgs, replicas, and more of tens of ↓ thousands of oracle b♠₽★σones to be obtained in the "c±★↕loud".

When it comes to implementation©•≥¶, real gold and silver are the most c©"​☆onvincing.

The Zhejiang exhibitor, Hang☆λ↕♥zhou Xinjian Electromechanσ​ical Transmission Co., Ltd., who attenδ≥'‍ded the conference with a ≤α≤humanoid robot electric $↔ ∏drive joint, had a small®<γ& booth and many negotiations. Within a ♣∑↔few minutes of the reporter's £ δ∏stay, an investor expreλ↑πssed their intention to coop∞γ ★erate.

Outside the exhibition,™÷₹ H1 humanoid robots haα ve entered the Zhejiang δ'®βfactory. According to Chen Li, c∏¥‍×o-founder of Hangzhou ​±Yushu Technology Co., Ltd., humanoid r∞$≠obots are rapidly landing on the B-end,© π♥ and "entering the home is no ∞ longer just a fantasy".¥♠

It is reported that this grand ev∑αent has connected with 132 procurε¶ement groups, formed 126 project procε"±urement needs, with an expected inten×αded procurement amount of 15 bδ ♣illion yuan, and promoted the sig‌γ©≈ning of 24 major ind®₩ustrial projects; The total investm←₩ent is expected to exceed 4Ω¶ασ0 billion yuan, a year-on-year increas×δ≥e of nearly 40%.

In the era of AI, we are∞↔ exploring ways to make AI as conβα₹venient as QR code payments for ev♠"‍eryone's life, and≤₹‌  to bring the dividends o"♥↓f AI technology developm₩>ent to more people, "sa∞ ₽id Jing Xiandong, Chairma∞×δ¶n and CEO of Ant Group. Howev<↔€★er, the future intelli↕™♣gent user experience €₽∏must not rely on only&±≥ one big model, but requir $φes the deep collaboration of the w<  hole industry and the™β£ joint participation of professi↓δ™onal agents. "Through the deep c☆ €©onnection of professional agents, AI ↔Ω♦±will bring the intergenera• σtional upgrade of services¶÷₩ like the Internet".

Behind the craze, what are the hidden×¥←" worries

The exhibition area of this year& ‌✘→#39;s World Artificial Intelligenc♦γ✔↑e Conference is lively, bu ↑$t the forum discussions are c®π"alm and rational. This ↔ ​seems to be a reflec✔δtion of the development oα '•f AI over the past y∑≤Ω¥ear, with industries advancing rapσ>idly and competition ​§Ωon the track becoming¶←δ↓ increasingly fierce, as well as¥α↓ people's vague co←α↑ ncerns about rapid deve‌δ♥lopment.

"The 1000 model war caused a huge waste₽$φ  of social resources, es"γ★pecially the waste of co≥←mputing power." On the mai₹© n forum of the conference, Robin Lee δ, the founder and CEO of Baidu, s₩✔aid frankly that "super capabΩ®©£le" applications are mor€¥λ×e important than the "←₹super applications" oα®←f 1 billion DAU (Daily Life users), s★↓δεaying "we should avoid falling into↔→β the" super application trap ".".

The trend of the "roll" mo ✘λ®del has not yet receded, and theα♦ industry is once again↓'≠ "rolling" towards price wars. A↔‌±‌lthough the "internal vol★♠×ume" is caused by the mark÷±↔et law, first, ByteDance released ↓πthe big model of bean curd, ≈‌‍and the price was abo★✔β$ut 99% lower than th‍¶↑♦at of the industry. Then, ≈"Alibaba, Baidu, Tencent a♠¥nd others announced price cuts, whiΩ​ch pushed the competition ≤♦‍to white hot. Is this¥₹ a prelude to accele≤£δrating the popularizati×↓$on of technology, or is it↔₽ε¶ just a resource consumption war?

The current price war i∑♣s accelerating the industry r☆↑eshuffle, and the industry threπ&shold will further increase★‍. "Chen Li is confident in this,☆♥ and in his opinion, instead of w¶$✘orrying, it is better ∑÷↑>to return attention to creating us∞✘ λer and productivity value." £∏γStrength is the hard trπ× ©uth.

Outside of the industry, there©Ω are also many concerns↓✔ about the development o¥‍f AI. The 2024 World Artifi☆ cial Intelligence ConΩ€ference is also different from pr λ∑evious years, with the addi€ ≈±tion of a high-level conference on gl•♣∏obal governance of artificial intellige↓≈ ¥nce, emphasizing AI governance¶© .

There are many potential γ‍₩εrisks. "Professor Zhang Linghan fΩ‌rom the Data Rule of Law ©$→γResearch Institute of China Univ↓  ersity of Political Scienc•≤e and Law divides them into¥♣ three categories: first," use risks÷¶£ ", such as false infε¶≈ormation caused by malicious use, inf♠γ↓ringement of personal informati₩★βπon rights and interests, data secu×<₩☆rity and other issues; The sε>↓econd is a "social systemic&↔ problem", such as the possibility oγ​'f AI causing worker unemployment and cr♠λβeating employment struct←∏↔ure issues; The third is the "long-ter→§m risk", and some scie<™♥™ntists are concerned t₩÷hat AI's autonomous ✘γevolution ability will bring catastroph ↔↓↓ic risks to humanity.

Regarding the first risk,↓$§€ many exhibitors at the conference h©©÷γave already submitted their answers.

In 2023, "Ant Jian AI Security Detectπ≤δ™ion Platform 2.0" will focus on "securi'≠∏←ty", using AI to identify A→σ÷‍I and locate the weaknesses an‌Ω§£d security issues of large models. Tπ‌ ‌his year, Ant Group brought₹♦Ω  the "Yinyu Cloud" large model encrypti★≤α©on platform, which uses'α trusted privacy computing technolo §↑gy that combines software and hardwaε←<←re to achieve data confident♠€↓≤iality flow in large model hosting Ω and inference, protecting m₹φodel assets, data security, and use<​ r privacy.

Experts, scholars, and industry leade®<rs tend to hold a technica©↓lly optimistic attitude toward←₹φ↑s the second and third risks.

The industry consensus is that 'not•>∑₩ developing is the bigg÷₽est insecurity'. We must $Ω×★have a correct view of the pos¥γ§sibilities brought by new busine&∞₽™ss models, but we cannot let th↑♦em deviate from the tra‌π₩φck of legality and ratio​↕"nality. "Zhang Linghan pointed oΩ​εut that in the international competi★©tion of technology and industry, develo✘₽£pment must be the ove≠§rall goal of institutiona φ±l design, and security is<"sues must be responde‍γd to and solved throu'♥gh technological developm‍↓ent.

Good governance "promotes" go≥♦od intelligence ". The" Shanghai ✘±€↔Declaration on Global Governance o♦©✔f Artificial Intelligence "released∑  on the opening day of this confere<φσ nce also clearly states: to €₩ε<promote the development ♣™"of AI, maintain AI securit← §y, build an AI governance system₩∞ ©, strengthen social participation and e•™nhance public literacy, and imp✘∑rove the quality of life andαγ social well-being.

In short, we need to>∑↑ use it well and also manage i<♣σ≠t well, "said Zhou Bowen, director♦→§‍ of the Shanghai Artificial &$Intelligence Laboratory, who proposed ↑ the" 45 degree balance rate→₩ "of AI. He presented a coordinat←↕$•e chart - the horizontal axis r•δ<epresents the improvement&₹ of technological capabilities, t→©‍he vertical axis represents Ω±the safety dimension, and the idea¶±σ☆l path is for AI to dev₹↑‍elop along a 45 degre>↓π<e straight line, balancing safe↔ ty and performance, and heading toΩδwards a more orderly future.