鍾蔚文、江靜之、陳百齡

Wei-Wen Chung, Chin-Chih Chiang, Pai-Lin Chen

人機協作、智能、契合、聚焦、創造、顯現

congruence, creation, Human-Machine Collaboration, intelligence, manifestation, orientation

本文探討在人類遇到AI之後所引發的智能（Intelligence）與知識體系危機，並提出針對性的對策。

全文分為三大部分：首先，將智能定義為與環境的契合能力，認為「人類與AI的相遇」是對這兩種智能系統契合度的重要試煉。這次相遇將決定它們在人類社會中的相對位置及角色，特別是在AI時代之後。其次，從智能的定義出發，分析自然智能（代表為人類）與人工智能（代表為AI）之間的本質差異。評估人類遇到AI之後，智能生態將如何變化。這一部分指出，雖然AI在表徵智能方面已經取得領先地位，但人類在聚焦、顯現和創造方面的智能仍然獨一無二，並且是AI所無法替代的。

最後，根據以上發現，以傳播學為案例，探討兩個議題。一、人如何在工作中與AI合作？以新聞工作為例，說明AI和人如何分工，以及專業人應該培養的能力。二、專業知識體系應如何調整和進化，以適應新的智能景觀？建議三個具體方向：（一）基礎研究，主張應從資訊處理轉向以具身之知、情境中之知為主軸之知識典範。（二）協物之學，探討並培養與物（如AI）合作分工之知識和智能，設計以人為本之知識架構和社會機制。（三）實務之學，強調在AI帶來各領域快速變遷的不確定年代，大學更應回歸大學應有之角色，以想像和設計未來之傳播實務為研究和教育之主軸。

Artificial intelligence (AI), as a general cognitive technology, significantly impacts nearly every aspect of modern life, especially in areas requiring cognitive tasks. Its widespread integration has presented unprecedented challenges to the adaptability of human intelligence and the relevance of established knowledge systems, particularly those based in universities. Through the lens of communication, this research explores these challenges, offering comprehensive analysis and proposing strategies for its acclimatization across four distinct sections.

Redefining Intelligence in the Age of AI

The first section of the paper proposes a theory of intelligence, which defines intelligence as the capacity of a species to align with its environment - a cornerstone of survival. Human intelligence is now being widely tested through the advent of AI. Encounters with AI can disrupt this harmony, necessitating an evolution in human intelligence that either reestablishes a balance or faces obsolescence. It is, however, important to note that the encounter is in essence of a test of congruence for both AI and humans. AI’s challenge is integrating into the human world in a way that complements human ecological norms. For human intelligence, the test is to distinguish itself from AI and navigate new paths of alignment in a landscape reshaped by AI’s influence.

The Congruence Test: Evaluating AI and Human Intelligence

The second section delves deeper into what is termed the congruence test. Starting with the premise that this test’s outcome for AI and humans hinges on the inherent characteristics of the two intelligence systems, the paper presents how natural human intelligence differs from artificial intelligence. Human intelligence is characterized by its embodied nature and its engagement with the world via a broad spectrum of strategies - representation, orientation, manifestation, and creation. In contrast, AI operates primarily through symbolic representations and lacks a physical embodiment, essentially limiting its modes of interaction.

This fundamental difference affects their performance in three key areas of the congruence test: coupling with the environment, sensing capabilities, and overall performance. Humans, being embodied and situated, can achieve optimal ecological coupling and are able to cover a broad spectrum of experiences. AI, while excelling in information processing, falls short in comprehensive sensing, highlighting that its representation abilities do not equate to genuine sensing. Compared to AI, which is adept at reckoning, humans are better at judgment and interpretation.
The congruence test also shows which human abilities will likely be rendered obsolete with the advent of AI and which might still be relevant. To be specific, human intelligence related to information processing tasks may pale compared to AI in terms of efficiency. Nevertheless, human intelligence characterized by embodied and situated knowing, may retain its uniqueness compared to AI.

How to Live and Work with AI?

Drawing from the insights of the congruence test, the following two sections discuss the implications for human intelligence and propose strategic adaptations.

The paper dwells upon two main tasks in the AI era. The first task is concerned with how humans live with AI. It suggests that the relationship between humans and AI is complementary rather than competitive. For instance, in journalism, a field heavily reliant on information processing, many tasks can now be relegated AI, including news gathering, analysis, content creation, and distribution. However, roles that require embodied knowledge, system oversight, and nuanced judgment remain firmly in the human domain.

To effectively collaborate with AI, professionals, particularly journalists, need to develop skills in computational thinking and critical data literacy that can enable them to leverage AI capabilities while maintaining the unique aspects of human intelligence. These skills are crucial as they are not easily replicated by AI systems, which lack the ability to fully understand or engage with the human world.

Reinventing Communication for an AI-Driven World

The final section of the paper argues for a significant reinvention of the field of communication to better align with the new, AI-driven ecology, giving special attention to three tasks.

First, as far as research is concerned, it calls for a shift from an information processing paradigm to one focused on embodied and extended knowledge. This shift requires broadening the foundational concepts and theories of the field to better suit the changing landscape.

Second, this paper emphasizes the historical co-evolution of humans with cognitive technologies and suggests that research and education should focus on developing a more collaborative and critical approach towards AI. This includes fostering a deep understanding of how to think and act with AI, critiquing technological developments, and establishing ethical guidelines for AI use.

Third, this paper advocates a futuristic orientation to practice education and research with emphasis on imagination and design. Universities are highlighted as critical institutions in this transition, tasked with imagining future trajectories and guiding practitioners through uncertain times. The urgency of this role increases as AI continues to disrupt traditional practices and assumptions, necessitating a proactive approach to curriculum and research that targets professional practice and the design of future communication strategies.

In conclusion, this paper not only addresses the immediate impacts of AI on professional intelligence, but also outlines a pathway for adaptation and coevolution, ensuring that human intelligence remains relevant and effective in an increasingly AI-integrated world.