來(lái)源：科技世代千高原
近年來(lái)，人工智能 (AI) 取得了令人矚目的進(jìn)步。語(yǔ)言模型能夠生成文本，圖像識(shí)別系統(tǒng)能夠創(chuàng)建逼真的視覺(jué)效果，機(jī)器能夠以驚人的速度掌握模式識(shí)別任務(wù)。然而，盡管取得了這些進(jìn)步，人工智能仍未達(dá)到人類智能真正卓越的水平：持續(xù)的終身學(xué)習(xí)以及從經(jīng)驗(yàn)中歸納總結(jié)的能力。這被稱為通用人工智能 (AGI)。
我的核心假設(shè)是，真正的通用人工智能 (AGI) 只有在人工智能能夠持續(xù)學(xué)習(xí)、靈活地實(shí)時(shí)調(diào)整其理解，而非僅僅依賴于大規(guī)模的一次性訓(xùn)練時(shí)才能實(shí)現(xiàn)。人類天生就會(huì)持續(xù)學(xué)習(xí)，根據(jù)與周圍環(huán)境的每一次新互動(dòng)來(lái)更新知識(shí)和理解。而目前的人工智能系統(tǒng)通常不具備這種能力。
計(jì)算機(jī)科學(xué)家、神經(jīng)科學(xué)家和工程師杰夫·霍金斯提出的“千腦”理論為實(shí)現(xiàn)這種持續(xù)學(xué)習(xí)提供了寶貴的見(jiàn)解?；艚鹚拐J(rèn)為，大腦是由眾多小型、分散的單元（稱為皮質(zhì)柱）組成的網(wǎng)絡(luò)。每個(gè)皮質(zhì)柱通過(guò)進(jìn)行預(yù)測(cè)、將其與實(shí)際感官輸入進(jìn)行比較并根據(jù)差異不斷更新，獨(dú)立地創(chuàng)建自己的“微型現(xiàn)實(shí)模型”。這種分散式結(jié)構(gòu)允許進(jìn)行穩(wěn)健、適應(yīng)性強(qiáng)且持續(xù)的學(xué)習(xí)，從而有效地避免了集中式神經(jīng)網(wǎng)絡(luò)方法中經(jīng)常遇到的災(zāi)難性遺忘。
為什么當(dāng)前的人工智能系統(tǒng)存在缺陷
如今，大多數(shù)人工智能系統(tǒng)嚴(yán)重依賴于大量的初始訓(xùn)練（預(yù)訓(xùn)練），之后則保持穩(wěn)定。與人類不同，這些系統(tǒng)無(wú)法持續(xù)適應(yīng)新情況或?qū)崟r(shí)吸收新信息。因此，人工智能系統(tǒng)通常難以將知識(shí)靈活地應(yīng)用于不可預(yù)見(jiàn)的任務(wù)或情境。
我的觀點(diǎn)是，AGI 的實(shí)現(xiàn)只有通過(guò)創(chuàng)造能夠在整個(gè)運(yùn)行生命周期內(nèi)持續(xù)學(xué)習(xí)、調(diào)整和保留知識(shí)的 AI 才能實(shí)現(xiàn)。受“千腦理論”啟發(fā)，實(shí)施一種去中心化的模塊化方法或許有助于解決這些問(wèn)題，因?yàn)樗梢宰?AI 動(dòng)態(tài)地整合新的經(jīng)驗(yàn)，同時(shí)保留先前學(xué)到的知識(shí)。
為什么參考框架對(duì)于真正的認(rèn)知至關(guān)重要
僅靠持續(xù)學(xué)習(xí)是不夠的。它需要一個(gè)關(guān)鍵要素：能夠整合感官輸入的穩(wěn)定參考框架。對(duì)人類而言，主要的參考框架是我們的身體。以識(shí)別咖啡杯為例：僅憑視覺(jué)識(shí)別是不夠的。只有當(dāng)我們親手觸摸它，感受它的形狀和重量時(shí)，我們才能真正理解并形成連貫的內(nèi)部表征。每一種感官輸入——視覺(jué)、觸覺(jué)和運(yùn)動(dòng)——都位于我們身體形態(tài)所提供的共享、穩(wěn)定的環(huán)境中。
人工智能要想發(fā)展出同樣復(fù)雜的認(rèn)知能力，還必須運(yùn)用清晰一致的參考框架。這些參考框架至關(guān)重要，因?yàn)樗鼈兪谷斯ぶ悄苣軌驅(qū)⒉煌母泄佥斎胝铣蛇B貫的心理表征，類似于人類通過(guò)身體解讀感官數(shù)據(jù)的方式。這種方法與世界模型的概念密切相關(guān)，人工智能首先需要深入理解并內(nèi)化各種對(duì)象和概念的特征和關(guān)系。只有創(chuàng)建了這種穩(wěn)定、集成的模型，人工智能才能有效地應(yīng)對(duì)全新的、前所未有的問(wèn)題。
運(yùn)動(dòng)技能和觸覺(jué)等復(fù)雜感官能夠顯著受益于真實(shí)的物理交互或高度逼真的虛擬模擬，它們能夠提供純虛擬輸入無(wú)法完全復(fù)制的關(guān)鍵情境。因此，這意味著，如果我們想要在人工智能中實(shí)現(xiàn)真正類似人類的認(rèn)知，就離不開(kāi)機(jī)器人技術(shù)；通過(guò)機(jī)器人系統(tǒng)或高度先進(jìn)的模擬技術(shù)，將實(shí)體化是邁向真正理解和通用智能的關(guān)鍵一步。
混合架構(gòu)方法
另一個(gè)懸而未決的問(wèn)題是，單靠去中心化架構(gòu)是否能夠完全實(shí)現(xiàn)持續(xù)學(xué)習(xí)，或者將去中心化和中心化元素相結(jié)合的混合架構(gòu)是否更有效。受“千腦理論”的啟發(fā)，我們可以想象無(wú)數(shù)個(gè)人工智能模塊，類似于大腦皮層柱，獨(dú)立學(xué)習(xí)并建模其局部感官輸入。同時(shí)，一個(gè)總體中央系統(tǒng)會(huì)將這些局部模型整合成一個(gè)統(tǒng)一的理解，在全球范圍內(nèi)協(xié)調(diào)響應(yīng)和決策。
這種混合方法可以在局部靈活性和全局一致性之間提供必要的平衡，為人工智能提供持續(xù)學(xué)習(xí)所需的穩(wěn)健性，而不會(huì)忘記過(guò)去的經(jīng)驗(yàn)。
結(jié)論與展望
實(shí)現(xiàn)通用人工智能可能需要從根本上轉(zhuǎn)向受人腦過(guò)程啟發(fā)的去中心化、持續(xù)學(xué)習(xí)模型。穩(wěn)定一致的參考框架，結(jié)合平衡去中心化局部學(xué)習(xí)和集中式全局協(xié)調(diào)的混合架構(gòu)，為實(shí)現(xiàn)通用人工智能 (AGI) 提供了充滿希望的途徑。在這些原則的指導(dǎo)下，未來(lái)的發(fā)展或許最終能夠彌合當(dāng)前的差距，使人工智能能夠真正像人類一樣思考和學(xué)習(xí)。
如今，人工智能系統(tǒng)已經(jīng)達(dá)到了成熟的水平，足以在組織內(nèi)部廣泛應(yīng)用——這不僅可以提高效率，還可以擴(kuò)展現(xiàn)有的商業(yè)模式，甚至創(chuàng)造前所未有的全新機(jī)遇，帶來(lái)巨大的附加值。事實(shí)上，如果真正的通用人工智能（AGI）需要更長(zhǎng)的時(shí)間才能出現(xiàn)，這對(duì)大多數(shù)公司來(lái)說(shuō)可能是有利的，因?yàn)樗赡軙?huì)迅速顛覆現(xiàn)有的商業(yè)模式。
在此之前，我建議各組織積極利用當(dāng)前的人工智能技術(shù)，尤其是基于代理的系統(tǒng)，來(lái)實(shí)現(xiàn)復(fù)雜工作流程的自動(dòng)化，并確保競(jìng)爭(zhēng)優(yōu)勢(shì)。理想情況下，他們應(yīng)該以創(chuàng)新的方式優(yōu)化和發(fā)展其商業(yè)模式，以至于即使有了通用人工智能 (AGI)，復(fù)制這些模式也會(huì)變得困難或缺乏經(jīng)濟(jì)吸引力。
福布斯技術(shù)委員會(huì)是一個(gè)僅限受邀者加入的社群，面向世界一流的首席信息官、首席技術(shù)官和技術(shù)高管。我有資格加入嗎？
在Twitter
上關(guān)注我
訪問(wèn)
我的
網(wǎng)站
Yusuf Sar, founder and CEO of
Hardwarewartung 24
. Pioneering in building sustainable data centers since 2001.
getty
Artificial intelligence (AI) has made remarkable strides in recent years. Language models generate texts, image recognition systems create photorealistic visuals, and machines master pattern recognition tasks at impressive speeds. However, despite these advancements, AI has yet to achieve what makes human intelligence truly remarkable: continuous, lifelong learning and the ability to generalize from experience. This is known as artificial general intelligence (AGI).
My central hypothesis is that true AGI can only be achieved when AI learns continuously, flexibly adapting its understanding in real time rather than relying solely on large-scale, one-time training sessions. Humans naturally engage in constant learning, updating their knowledge and understanding based on every new interaction with their surroundings. Current AI systems typically do not possess this capability.
The "Thousand Brains" Theory proposed by Jeff Hawkins—computer scientist, neuroscientist and engineer—provides valuable insights into achieving this kind of continuous learning. According to Hawkins, the brain operates as a network of numerous small, decentralized units called cortical columns. Each column independently creates its own "miniature model" of reality by making predictions, comparing them to actual sensory inputs and continually updating based on discrepancies. The decentralized structure allows robust, adaptable and continuous learning, effectively preventing the catastrophic forgetting frequently encountered by centralized neural network approaches.
Why Current AI Systems Fall Short
Today, most AI systems rely heavily on extensive initial training (pre-training) and remain static afterward. Unlike humans, these systems do not adapt continuously to new situations or incorporate new information in real time. Consequently, AI systems often struggle to apply knowledge flexibly to unforeseen tasks or contexts.
My argument is that AGI can only be achieved by creating AI that can continuously learn, adjust and retain knowledge throughout its operational lifetime. Implementing a decentralized, modular approach inspired by the Thousand Brains Theory might help solve these issues by allowing AI to dynamically integrate new experiences while preserving previously learned knowledge.
Why Reference Frames Are Essential For True Cognition
Continuous learning alone is insufficient. It requires a crucial component: stable reference frames that integrate sensory inputs. For humans, the primary reference frame is our body. Consider recognizing a coffee cup: Visually identifying it alone is incomplete. Only when we physically touch it, feeling its shape and weight, can we truly understand and form a coherent internal representation. Each sensory input—visual, tactile and motor—is positioned within a shared, stable context provided by our physical form.
For AI to develop similarly sophisticated cognitive abilities, it must also employ clear and consistent reference frames. These reference frames are essential because they enable AI to integrate diverse sensory inputs into coherent mental representations, similar to how humans interpret sensory data through their bodies. This approach is closely linked to the concept of world models, where an AI first needs to deeply understand and internalize the characteristics and relationships of various objects and concepts. Only after creating such stable, integrated models can AI effectively tackle completely novel, previously unseen problems.
Complex senses like motor skills and haptics significantly benefit from actual physical interaction or highly realistic virtual simulations, providing critical context that purely virtual inputs may not fully replicate. Consequently, this implies we cannot bypass robotics if we aim to achieve truly human-like cognition in AI; physical embodiment, through robotic systems or highly advanced simulations, is an essential step toward developing genuine understanding and general intelligence.
A Hybrid Architectural Approach
Another open question is whether decentralized architectures alone can fully realize continuous learning or if a hybrid structure, combining decentralized and centralized elements, might be more effective. Drawing inspiration from the Thousand Brains Theory, one can imagine numerous AI modules, analogous to cortical columns, independently learning and modeling their local sensory inputs. Simultaneously, an overarching central system would consolidate these localized models into a cohesive understanding, coordinating responses and decisions on a global scale.
This hybrid approach could offer the necessary balance between local flexibility and global coherence, providing AI with the robustness required to continuously learn without forgetting past experiences.
Conclusion And Outlook
Realizing artificial general intelligence will likely demand a fundamental shift toward decentralized, continuous learning models inspired by human brain processes. Stable and coherent reference frames, combined with hybrid architectures balancing decentralized local learning and centralized global coordination, offer promising pathways toward AGI. Future developments guided by these principles might ultimately bridge the current gap, enabling AI to genuinely think and learn like a human.
Today's AI systems have already reached a maturity level sufficient for broad adoption within organizations—not just to increase efficiency but also to expand existing business models or even create entirely new opportunities that were previously unattainable, delivering tremendous added value. In fact, it could be beneficial for most companies if true AGI takes more time to emerge, as it might rapidly disrupt established business models.
Until then, I suggest organizations proactively leverage current AI technologies, particularly agent-based systems, to automate complex workflows and secure competitive advantages. Ideally, they should optimize and evolve their business models in such innovative ways that replicating them, even with AGI, becomes challenging or economically unattractive.
Forbes Technology Council is an invitation-only community for world-class CIOs, CTOs and technology executives. Do I qualify?
Follow me on Twitter or LinkedIn. Check out my website.
閱讀最新前沿科技趨勢(shì)報(bào)告，請(qǐng)?jiān)L問(wèn)歐米伽研究所的“未來(lái)知識(shí)庫(kù)”
https://wx.zsxq.com/group/454854145828
未來(lái)知識(shí)庫(kù)是“ 歐米伽 未來(lái)研究所”建立的在線知識(shí)庫(kù)平臺(tái)，收藏的資料范圍包括人工智能、腦科學(xué)、互聯(lián)網(wǎng)、超級(jí)智能，數(shù)智大腦、能源、軍事、經(jīng)濟(jì)、人類風(fēng)險(xiǎn)等等領(lǐng)域的前沿進(jìn)展與未來(lái)趨勢(shì)。目前擁有超過(guò)8000篇重要資料。每周更新不少于100篇世界范圍最新研究資料。 歡迎掃描二維碼或訪問(wèn)https://wx.zsxq.com/group/454854145828進(jìn)入。
截止到3月31日 ”未來(lái)知識(shí)庫(kù)”精選的百部前沿科技趨勢(shì)報(bào)告
（加入未來(lái)知識(shí)庫(kù)，全部資料免費(fèi)閱讀和下載）
熱門跟貼