What Was Observed? (Introduction)

• The study explores the idea that cognition and sentience, or the ability to experience, can exist in many types of systems, not just those with brains.
• It challenges the old belief that mental functions must be tied to specific brain structures and shows that diverse systems, including non-neural systems like plants or engineered robots, can also exhibit cognitive abilities.
• The idea is that cognition can be realized in different substrates (or mediums), like non-living materials or synthetic entities like cyborgs, which might act intelligently even if they don’t have a biological brain.

What is Sentience and Cognition?

• Sentience refers to the capacity to experience subjective thoughts and feelings, like pain or joy.
• Cognition refers to mental functions like thinking, learning, decision-making, and adapting to new information.
• Traditionally, scientists believed only organisms with complex brains could display cognition or sentience.
• Recent discoveries show that even simpler systems, such as single cells, plants, or even robots, can display forms of cognition and decision-making.

How Can Cognition Be Realized in Different Systems? (Multiple Realizability)

• Cognition is “multiply realizable,” which means it can appear in many different forms across various systems and substrates (materials or biological components).
• For example, both machines and biological organisms can perform computations (like solving problems or making decisions) but with different mechanisms.
• Some non-neural organisms (like slime molds or plants) show behaviors that suggest they can learn, adapt, and solve problems, even without a brain or nervous system.
• This suggests that cognition does not require a brain and can emerge in different kinds of systems with the right properties for processing information.

Can We Expand the Definition of Cognition Beyond Biological Systems?

• The paper argues that cognition is not limited to living systems. It can also be realized in bio-engineered systems or synthetic intelligence (AI).
• These systems could include cyborgs (a mix of biological and mechanical components), robots with artificial intelligence, and even materials that can process information.
• The challenge is to identify what kinds of elements are necessary to create systems that can process information, learn, and adapt.
• Instead of asking about which specific system (like a brain) is required for cognition, we should focus on the abstract elements needed to create any cognitive system.

What Is the Difference Between Adaptation and Learning?

• Learning typically refers to a system’s ability to change based on previous experiences or inputs. This can be seen in animals, cells, and even machines.
• However, similar processes in simple organisms or non-living materials are often called “adaptation” instead of learning, despite being fundamentally the same.
• The paper argues that these arbitrary distinctions should be removed, allowing us to view all systems that exhibit similar behaviors as performing cognitive functions, regardless of their composition (biological or mechanical).

Sentience in Non-Neural Systems (What Is It Like to Be Something Else?)

• Sentience, or the ability to experience things, is a private process that cannot be directly observed or measured. We infer it by looking at behaviors (like how something moves or reacts).
• The challenge is that many systems (like single cells or robots) might display intelligent behaviors, but we can’t be sure if they are “experiencing” anything.
• Humans often assume other beings are sentient if they show similar behaviors to us. However, the paper warns that this approach might miss sentience in systems that behave differently, like non-human animals or artificial intelligence.

What Are the Minimal Requirements for Sentience? (Complexity and Scale)

• The paper asks whether sentience can exist in simpler systems, like individual cells or single neurons, or if it only emerges when many cells work together in complex organisms.
• While individual cells show behaviors like decision-making and learning, it’s still unclear if they experience subjective states like humans do.
• The paper suggests that we might need to rethink the scale and complexity required for sentience, recognizing that even small systems might have some form of experience.

How Can Bioengineered Systems Help Us Understand Cognition?

• Advances in bioengineering, like creating hybrid robots (called “hybrots”), have opened new avenues for studying cognition in non-biological systems.
• Hybrots involve biological cells controlling robots, allowing scientists to test how cells respond to sensory input and how these responses can lead to intelligent behavior.
• Bioengineering also allows us to create modular neural circuits in the lab to isolate and study specific cognitive functions, which can give insights into how cognition might arise in other types of systems.

Ethical Considerations (New Ethical Frameworks)

• As our understanding of cognition broadens, we will need new ethical frameworks to consider systems that may not share our biological makeup or evolutionary history, such as cyborgs, synthetic intelligences, or bioengineered beings.
• Traditional distinctions between “sentient” beings (like animals) and “mechanical” systems (like machines) are no longer sufficient. These old labels are becoming outdated as more diverse types of cognitive systems are created.
• The ethical challenge is to develop ways to treat all systems fairly and ethically, regardless of whether they are made of biological material or synthetic components.

Key Conclusions (Discussion)

• The concept of cognition is far more expansive than previously thought. It is not limited to brains and can emerge in a variety of systems, including non-neural and synthetic ones.
• Sentience is likely more widespread than we realize, and we must develop new ways to detect and interact with sentient systems that do not fit traditional categories.
• As technology progresses, we will need to reconsider old ethical frameworks and develop new approaches to address the ethical implications of interacting with diverse forms of sentience.

观察到的是什么？ (引言)

• 本研究探讨了认知和感知的概念，即体验的能力，可以在多种不同的系统中存在，而不仅仅是大脑。
• 它挑战了将所有心理功能与大脑特定结构相关联的传统观点，显示了包括非神经系统（如植物或工程机器人）在内的多种系统也可以表现出认知能力。
• 这一观点认为，认知可以在不同的基质（或介质）中实现，像非生物材料或合成实体一样，它们即使没有生物学的大脑，仍然可能表现出智能行为。

什么是感知和认知？

• 感知指的是体验主观思想和感受的能力，如痛苦或喜悦。
• 认知指的是思考、学习、决策和适应新信息等心理功能。
• 传统上，科学家认为只有具有复杂大脑的生物才能展示认知或感知。
• 最近的发现表明，即使是更简单的系统，如单细胞、植物，甚至机器人，也能表现出某种形式的认知。

如何在不同的系统中实现认知？ (多重实现性)

• 认知是“多重实现的”，意味着它可以以多种不同的方式出现在各种系统和基质中（材料或生物成分）。
• 例如，机器和生物体都可以执行计算（如解决问题或做决策），但机制不同。
• 一些非神经生物体（如变形虫或植物）表现出可以学习、适应和解决问题的行为，即使没有大脑或神经系统。
• 这表明，认知并不需要大脑，而是可以在具备处理信息能力的其他系统中出现。

我们能否将认知的定义扩展到生物系统之外？

• 文章认为，认知不仅限于生物系统，它也可以在生物工程系统或合成智能（AI）中实现。
• 这些系统可以包括赛博格（生物和机械组件的混合体）、具有人工智能的机器人，甚至可以处理信息的材料。
• 挑战是确定创建能够处理信息、学习和适应的系统所需的基本元素。
• 与其问认知需要哪种特定的系统（如大脑），不如关注创建任何认知系统所需的抽象元素。

适应与学习的区别是什么？

• 学习通常指系统根据以前的经验或输入做出变化。这可以在动物、细胞甚至机器中看到。
• 然而，在简单的生物体或非生物材料中，类似的过程通常被称为“适应”，尽管它们本质上是相同的。
• 本文认为，这些人为的区分应当被消除，允许我们将所有表现出类似行为的系统视为执行认知功能，而不管它们的组成是生物的还是机械的。

非神经系统中的感知（像其他事物一样存在是什么样的？）

• 感知，或体验的能力，是一种私密的过程，不能直接观察或测量。我们通过观察行为（例如如何运动或反应）来推断它。
• 挑战在于，许多系统（如单细胞或机器人）可能表现出智能行为，但我们无法确定它们是否“体验”了任何东西。
• 人类通常假设其他生物是有感知的，尤其是当它们表现出与我们相似的行为时。然而，本文警告，这种方法可能会错过那些表现不同的系统中的感知，诸如非人类动物或人工智能。

感知的最小要求是什么？（复杂性和尺度）

• 文章探讨了感知是否只存在于更复杂的系统中，如多细胞生物，还是它是细胞层次的基本属性，只是在与人类感知相符的时间尺度和空间尺度下才显现出来。
• 虽然个别细胞表现出像决策、学习这样的行为，但目前尚不清楚它们是否经历了类似人类的主观状态。
• 文章认为，我们可能需要重新思考感知所需的尺度和复杂性，认识到即使是小系统也可能拥有某种形式的体验。

如何通过生物工程系统帮助理解认知？

• 生物工程的进展，如创建混合机器人（“赛博机器人”），为研究非生物系统中的认知开辟了新的途径。
• 赛博机器人涉及生物细胞控制机器人，使科学家可以测试细胞如何响应感官输入，以及这些反应如何导致智能行为。
• 生物工程还使我们能够在实验室中创建模块化神经电路，逐步隔离和研究特定的认知功能，这可以为我们理解认知如何在其他类型的系统中产生提供见解。

伦理考虑（新的伦理框架）

• 随着我们对认知理解的拓展，我们需要为考虑那些没有与我们相同进化谱系、组成或来源的存在，制定新的伦理框架，诸如赛博格、合成智能或生物工程存在。
• 传统上，“感知”和“机械”系统之间的区别不再足够。当更多类型的认知系统被创造出来时，这些旧有的标签逐渐变得过时。
• 伦理挑战在于，制定公平和道德的方式来对待所有系统，不管它们是由生物材料还是合成成分构成。

关键结论（讨论）

• 认知的概念比以前认为的要广泛得多。它不仅限于大脑，可以在各种系统中出现，包括非神经系统和合成系统。
• 感知可能比我们意识到的更加普遍，我们必须开发新的方法来检测并与不符合传统范畴的感知系统互动。
• 随着技术的进步，我们需要重新思考旧有的伦理框架，并制定新的方法来解决与各种形式感知互动的伦理问题。