EU Artificial Intelligence Strategy and Prospects for China-EU AI Governance Cooperation

W impact of artificial intelligence （AI） on national security，economy，and politics is becoming increasingly evident. Majorcountriesare formulating national AI strategies to compete for future technological dominance. Against thisbackdrop， the European Union （EU） has chosen a path centered on norm-shaping and values leadership， attempting to dominate global AI governance through rule-setting and ethical initiatives. However， this strategy of \"regulation first” faces increasingly prominenttensionsand imbalances inthe face of rapidly evolving technological innovation and intensifying geopolitical competition. To enhance itstechnological competitiveness whilemaintaining itsadvantagein norms， the EU is attempting to adjust its AI strategy，which will bring opportunitiesforthedevelopmentofAIin the EU and China-EU AI governance cooperation.

THECOREPILLARSOFTHEEU'SAI STRATEGY

In 2018， the European Commission issued thepolicydocumentArtificial Intelligence for Europe， explicitly incorporating AI into the policy agenda for the first time. It proposed strengthening Ramp;D investment， improving data infrastructure， and establishing ethical norms to build a“trustworthy AI\"system.In2020，theEUfurther releasedtheWhitePaperonArtificial Intelligence，proposing a dual-pillar model centered on“excellenceand trust\"，signaling a shift in itsregulatoryagenda fromprinciple orientation to institutional design. In April 2021， theEuropean Commission formally unveiledadraftfortheEUArtificial Intelligence Act， establishing ariskbased governance framework，marking the EU's AI strategy transition froma value-oriented“soft law”approachtoalegallybinding“hard law” regulatory model. On 1 August 2024， the EU Artificial Intelligence Act completeditslegislativeprocessandenteredinto force，becomingtheworld's firstcomprehensive legal framework for regulating AI. Overall， the EU'sAI strategyhasgraduallyformed three corepillars：norm-shaping，leadership in rule-making，and technological sovereignty.

First，norm-shaping.This involves guiding the direction of technology applicationthroughavalue-oriented regulatory framework，establishinga strictrisk-categorization based regulatorymodel via the AI Act， and providinga“Europeanparadigm”forglobalAI governance.Asasupranational internationalorganization， theEUexcels at amplifying its influence through rule-making，knownas the“Brussels Effect\".InthefieldofAI，theEUalso attemptstosetaglobal benchmark withastringentregulatory framework. The AI Act constructs a risk-categorization-based regulatory system covering high-risk system scrutiny，transparencyrequirements， datagovernance， and compliance mechanisms， forming anend-to-endregulatorychainfrom designto deployment.ThisActnot onlyprovidesaunified governance basisforEU member states butalso sets standardthresholdsforglobal AIcompaniesenteringtheEuropeanmarket， therebypromotingtheinternational diffusion and institutional transplantation of the EU governance paradigm.

Second， leadership in rule-making. TheEU seeks to export its“trustworthy\"and“human-centric\"valuesinternationally， vying for rule-making powerand influence inmultilateral mechanisms. During the 2023 G7 HiroshimaSummit，theEUcollaborated with other members to promote the Hiroshima AI Process， emphasizing frontierregulation，risk assessment， andtransparentaccountabilitymechanismsforgenerativeAI.TheEUalso actively supports the UN SecretaryGeneral's proposal to establisha“global AI governance body\"， suggesting thatUNESCO lead thedevelopment ofglobalAI ethical standards.TheEU is also building AI capacity-building mechanisms with Global South countriesthroughinitiativeslike theAfricaEurope Digital Economy Partnership and the ASEAN-EU Strategic Partnership，promoting European governance experience.Thisreflectsthe EU'sattempt to carve out anew governance approach beyond the U.S.and China， leveraging its advantage in norms to expand strategic depth.

Third， technological sovereignty. This emphasizes enhancing digital technology autonomy and reducing dependence on external technologies. Compared to the U.S.and China， the EU's strength in AI technology Ramp;D and industrial innovation is significantly weaker， causing a strategic imbalance. On the one hand，few European companies rank among the AI frontier， highlighting relative lagging behind in Ramp;D and application capabilities.Despite efforts like the Digital Europe Programme launched in 2025 and innovation schemes supporting startups and SMEs released in 2024， Europe's overall innovation ecosystem still struggles to match U.S.Silicon Valley or China'snewAI industry clusters represented by Hangzhou and Shenzhen. On the other hand， European investmentin AI is far behind that of the U.S.and China. Statistics show that in 2O23，total AI venture capital in the EU was about USD 8 billion， compared to USD 68 billion in the U.S.andUSD15billioninChina.Due to these technological and industrial shortcomings，the EU findsit difficult to keep pace with the U.S.and China in the global AI race. Consequently， European Commission President Ursula von der Leyen stated at the 2O25 Paris AIAction Summit that Europe should not blindly follow others but leverage its strengths in research and talent to build a“European brand\"of AI.

THEIMBALANCEDILEMMAOFTHE EU'SAI STRATEGY

Although the EU has constructed an AI governance frameworkwith strategicvision，itsimplementation process reveals multiple imbalances essentially manifested in the structural tension between its governance systemand the laws of technological innovation.From strategic orientation to institutional design，from technological development tomarket transformation，and from member state coordination to global rules competition，these imbalances expose the EU'sinternal structural contradictions in resourcesand policies， andseverelyconstraintheachievement of its strategic goals.

First，the lack of military support. Former U.S. Secretary of State Henry Kissinger ranked nuclear weapons， cyber-security，andAI ashistorically significantstrategic technologies，emphasizing theirdeep embedding within national security strategies—not merely economic-technical issues，but matters relating to military competition and international security stability. The EU's current AI strategy primarily focuses on \"soft issues”like ethical norms，risk control，andhumanrightsprotection，while neglecting the integrated planning and civil-military fusion pathwaysrequired forAI asa corenational strategic re source.Thisleavesthe EUsignificantly ill-equipped to address the security challenges posed byAI asa strategic technology—unable to support the authority of itsgovernance framework through military force or intelligence systems， and struggling to play an effective role in global AI governance.

Second，the“over-securitization\" tendency in technologygovernance and innovation suppression. The EU's AI strategy，by overly emphasizing po litically-led security-first tactics，exhibits adegree of\"over-securitization\"tendency，overlooking the inherent structural propertiesofAI technology. Forrapidly iterating strategic technologies，excessive investment in regulatory resources inhibits technological breakthroughs Thecomplexriskmanagement，data governance，and compliance systems builtbytheAIActfail to fullyaccount for the uncertainty of AI technology evolution and the speed of iterative innovation.Its static riskclassification framework struggles to adapt to dynamicinnovation pace.The EU'scurrent focus on control and ethics，while theoretically strengthening technologicalsecurity，practicallyleadstodelays in core technology breakthroughs and application transformation.This“security-first\"regulatory paradigm essentially ignores the intrinsic developmental logic of AI as a disruptive technology， creating a goal misalignment between institutional innovation and technologi cal evolution.

Third，thestructuralweaknessesin industrial ecosystem and systemicinnovationdeficiency.TheEU'straditional industrial strengthslie in manufacturing， energy，etc.，while support for emergingtechnologyindustrieslike the internet and AI is relatively lagging behind.A deep-seated reasonisthecautiousattitude of European capital markets towards high-risk tech investments， making itdifficult forstartups to secure large-scale financing.Furthermore， the dual governance structure between the EU and its member states significantly constrainstheEU'sabilityto formulate cohesive industrial policiesand build an innovation ecosystem.Consequently， despite Europe'sstrong foundation in researchtalentandpatentreserves，it has failed to forma competitive industrialecosysteminareaslikeventure capital，businessmodelinnovation，and technology commercialization. This \"strong in research，weak in industry\" paradoxreflectsthestructural imbalanceintheEU'sinnovationsystem.

Fourth，thedilemmabetween strategicautonomy aspiration and realisticdependenceininternational cooperation. In AI international cooperation，the EU'saspiration forstrategic autonomyclasheswith itsrealityof de pendence on theU.S. On the one hand theEU advocates the concept of\"digital sovereignty\"， striving fora leading role in global AI governance while avoiding excessivedependenceonanymajor country. On the other hand， driven by geopolitical and securityconsiderations， itsactual policies often lean towards joiningU.S.-led“smallcircles”likethe G7. This camp-based tendency exacerbates the fragmentation of global AI governance and conflicts intrinsically with the multilateralism principles championed by the EU. When engaging in AI cooperation with China， the EUfacesboth autonomy concerns and pressure fromtheU.S.Itspoliticization and instrumentalization of technological competition not only fail to enhance security effectiveness but also damage the open and cooperative international innovation environment， highlighting thecontradiction betweentheEU's autonomy and dependence in global technological competition.

POTENTIAL ANDCONSTRAINTSOF CHINA-EUCOOPERATIONUNDERAI STRATEGICTRANSFORMATION

TheEUiscurrentlyatacritical juncture in its AI strategic transformation.Againstthebackdropof“excess regulation，insufficient competitiveness\"， the EU attempts to continue leading the global agenda on ethics and rules through the AI Act while also recognizing its systemic weaknesses in enterprise numbers， investment capacity， and infrastructure， urgently needing external cooperation to break the deadlock.New pathways offered by Chinese companies likeDeepSeek，characterized bylow cost and high efficiency，present a practical opportunity for cooperation. Facing insufficient local AI industry investment and strained computing resources， countries like Germany and France have begun referencing DeepSeek's development path， re-evaluating their own technological layouts. Its conceptsof“decentralization”and “resource adaptation” provide a practical model for Europe to explore an AI development path suited to its own endowments， also expanding the space for technological complementarityin China-EU cooperation. Specifically， the prospects for China-EU AI cooperation can be observed from three dimensions： technological Ramp;D complementarity， sustainable development，and global governance.

In terms of technological Ramp;D， the EU and China possess complementarycomparativeadvantages.The EUis strongin fundamental research andunderlying algorithms，but its research strengthshave not translated intolarge-scalecommercial applicationsand industrial competitiveness. China excels uniquely in AI engineering and large-scale application， cultivating companies like DeepSeek withstrong industryadaptability，high scenariosensitivity，and system-level technologyintegration capabilities. Thecomplementaritybetween the EU's“theory-driveninnovation\"model and China's“application-driven inno vation”modellaysthefoundationfor building and deepening technological and governance cooperation between the two sides.

In terms of sustainable development，Chinaand theEUsharecommoninterests.BoththeGreenDeal proposed bytheEUandtheinnovation-driven development strategy advocated by China require AI support. Furthermore， facing common sustainabledevelopmentchallengessuchas healthcare，aging populations， climate change，and energy transition，both ChinaandtheEUhavetheintentand capabilityto leverageAI to improve public service levels， optimize social se curitysystems，and enhance industrial energyefficiency.Itisevident that ChinaandtheEUsharebroadcommon interestsin promotingsustainable development，providing fundamental conditions for jointly formulating morerobustinternationalrules.

In terms of global AI governance， ChinaandtheEUholdsimilarpositions on upholdingmultilateralism and opposing unilateral， instrumentalized governance. Current global AI governance faces numerous disruptions from geopolitical competition. Comparedto the U.S.'sunilateral measuresandexportcontrolscharacterizedbya“small yardwithhigh fences”approach，ChinaandtheEU aremoreinclinedtorelyonmultilateralplatformslike theUNandG20 tomaintainopenandinclusive international governance，building rule systemsthatbalancethe interestsof all parties.This shared stance provides significant impetus for deepening bilateralcooperation.

Nevertheless，theEUstill harbors practical reservations when evaluating the technological solutionsofChinese AI enterprises represented by DeepSeek. The EU’s AI ecosystem has long beendeeply embedded withintheU.S.- dominated technology supply chain system，inevitablyleading to conservatismandcautionregardingdata privacy compliance， national security considerations，and ethical values.

On data privacy，the EU has built the world'sstrictest personal dataprotection frameworkcenteredon the General Data Protection Regulation （GDPR）， emphasizing user privacyand data localization. Therefore， any AI product entering the European market （like DeepSeek） mustcomplywithGDPRand similar regulations.The EU is concerned about whetherthesemodelswilladoptits high-standard privacy protocols， such as transparency in training data collection and user data processing，obtaining user consent， and ensuring that European dataisnottransmittedabroadwithout authorization.

On national security， the EU often finds itselfcaughtinthestructural contradictionbetweenits“aspirationforstrategicautonomy”andits \"transatlantic securitydependence\" in its China AI policy. This manifests particularlyin areaslike technical standard setting and enterprise scrutiny，showinga risk preference trending towards politicization， leading to increasingly stringent security reviews of Chinese tech companies in recentyears.Ontopof this，European intelligence agencies and policymakersworryabout Chinese companies collaboratingwith their government， potentiallyembeddingsecurityrisks within their products.This suspicion causes Europe to scrutinize Chinese AI products like DeepSeek notonly from acommercial perspective butalsowith heightened vigilance fromanational securityangle.

On ethicsand values，European publicopinionoftencites China's “social creditsystem\"asanexample， questioningwhethertheethical orientation of Chinese tech products aligns withEuropean values.China's social credit pilots have been misrepresented bysomeWesternmediaassymbolsof \"big data totalitarianism”，raising concernsamongtheEuropeanpublicthat China uses AI technology for social control， impacting the ethical principles Europe upholds.

PATHWAYSFORCHINA-EU AIGOVERNANCECOOPERATION

Given the significant common interestsand structural complementaritiesbetweenChinaandtheEUin the field of AI governance， both sides possessarealisticfoundationand strategic necessity for building deep collaborative relations. Strengthening cooperation can not only effectively preventtheadversarialdeterioration ofdigital geopoliticsbutalso inject stabilizing forces into building an open and inclusive global AI governance system.Buildinga China-EU cooperation framework requiressynergistic advancementacrossmultipledimensions：institutional dialogue， standard coordination，multilateral governance， and enterprisecooperation.

First，establishingahigh-level China-EU dialogue mechanism on AI governance. Chinaand the EU should establish institutional exchange platforms on digital policy and AI governance， settingupadedicated joint working group on AI governance. Withinthisframework，thetwosides canengageincandidcommunication onspecific issues like data protection， algorithm transparency，andAI compliancecertificationtoenhancemutual trust. Focusing on AI governance， China and the EU can open direct communication channels and regularlyconvey their consensus to the world through multilateral platforms. The two sides can also encourage in teractions among legislative bodies， regulatoryagencies，enterprises，and academia，promoting the formationof a composite dialogue system integrating\"policy-technology-industry\".

Second，promoting the alignment and joint development of AI governance standards.While the EU implements itsAI Act， China isalso accelerating the improvement ofregulations concerning algorithm security， data outbound transfer，and security assessments. China and the EU should seize this opportunity to coordinate on technical standards and governance rules. On the one hand， they can encourage cooperation between their standard-setting bodiesand research groups to exchange views on AI ethical guidelines，risk classification，testing and evaluation methods， striving for compatibility between their respective standardsto reduce corporate compliance conflicts across the two markets. On the otherhand， China and the EU can provide more public goods within the International Organization for Standardization and other global multilateral forums， jointly proposing international standards for AI governance，promoting the formation of global standard consensus， providing accessible technical reference frameworks for developing countries， and breaking down the“technical threshold barriers” of AI governance. Both sides can also pilot mutual recognition mechanisms for third-party assessment systems，reducing the“double costs\" companies face when complying in both Chinese and EU markets， and enhancing the efficiency of crossborderdeplovmentforAI products

Third， strengthening cooperation and coordinationof positionsin multilateral platforms.Askeyparticipants in global governance，China and the EUhave a responsibility to lead inclusive global AI governance. Both sides should join hands to leverageplatforms like theUN and G2O to promote anopen and cooperative governance model. At the UN level， China and the EU can support the work of the HighLevel Advisory Body on AI and promote inclusive governance based on broad representation.Both sidescan also sustain collaboration within multilateral organizationsand mechanisms like the UN Internet Governance Forumand UNESCO to resist the impact of the politicization of technological issues on the global governance order. AttheG2O level， China and the EU should uphold the spirit of the 2019 “G20 AI Principles” and further build consensus on issues like AI regulatory sandboxes and cross-border data flowrules.Against the backdrop of an increasingly“geopoliticized” trend in global technology development， the U.S.，by establishing“small circles\"and promoting regulatory standard alliances， is instrumentalizing the governance of key areas like AI. The global technology governance system is showing tendenciesofcamp formation and ideological division. As key forces in maintaining the global order for AI， China and the EU should jointly resist the camp-based trend in technology governance，offer theworlda third institutional choice，and help global AI governance move beyond the structural dilemma of “strong powers setting rules-others passively accepting\"，thus avoiding the fragmentation of global technical rules.

Fourth， building mechanisms for China-EUAIenterprisecollaboration and risk response. Beyond intergovernmental cooperation， promoting mutual trustandcollaborationmechanisms between enterprises is equally important.China and the EUcan encourage leading AI enterprises to establish joint Ramp;D platforms， sharing certain trainingresourcesand evaluation tools to enhance the cross-context and cross-cultural adaptability of models.Furthermore， exchanges between AIenterprises，platformcompanies， and standard-setting bodies should be promoted to bridge the practical gapbetweenpolicy formulation and technology implementation. In building consensus on technological risks， effortsshouldbemadeto establish an AI Risk Assessment Center jointly supported by Chinaand the EU，centrally conducting ethical assessments， security vulnerability testing，and misuse warnings for large models and generativeAI，enhancingboth sides' shared understanding of underlying AI risks. Thisdeep integration of enterprisesand governancemechanisms canserveasaninstitutional fulcrum for China-EU cooperation and provide amodelof\"state-enterprise\"collaborativegovernance for the world.

In conclusion， the EU's AI strategy reflectstheprofound tension between its aspirations for global rule-shaping and technological sovereignty. Facing the paradigm shifts brought by emerging large model technologies， no singleactorcanindependentlyaddress the complex challenges of AI governance. On the occasion of the 50th anniversaryofChina-EUdiplomatic relations，both sides need to uphold the concept of win-win cooperation and build strategic mutual trust in areas such as AI ethical standards， security certification，and cross-border collaboration. Only by adhering to the spirit of multilateralism and promoting the formation of a global AI governance system characterized by “technological inclusiveness，rules coordination， and shared responsibility\"can technological innovationand social risksbe effectivelybalanced，soastoachieve sustainable development in the AI era. This collaboration concerns not only the strategic interests of China and the EU themselves but isalsoacrucial path to shaping the global technological order of the 21st century.