摘要:
工智能(AI)的发展被大致分为了弱AI (专用AI)和强AI (具身或通用AI),似乎缺乏中间形态或阶段。从"魅"的视角审视,AI可分为"祛魅" "返魅"和"拟魅"三种形态或阶段。"祛魅的AI"是指"无心"的机器智能,纯粹展现工具理性;"返魅的AI"是指"有心"的机器意识系统,旨在实现具身心智的通用智能;"拟魅的AI"是处于"无心"和"有心"之间某一"中间状态"的"拟心"状态,目的是彰显交互意识,完成各种认知任务。这种划分有利于新一代AI的健康发展,不仅可避免两种智能因进化产生的"解释鸿沟",也可避免技术壁垒和伦理困境。
Abstract:
The development of Artificial Intelligence (AI) can be roughly divided into weak AI (specialized AI) and strong AI (embodied or AGI), and there seems to be a lack of intermediate forms or stages. From the perspective of “charm” or “phantom”, AI can be categorized into three forms or stages, namely, “dispelling charm”, “returning to charm” and “imitating charm” stages. The “AI that dispelling charm” refers to “mindless” machine intelligence, purely demonstrating instrumental rationality; the “AI that returning to charm” refers to “mindful” machine consciousness systems, aiming at realizing a general intelligence with embodied mind; “AI that imitating charm” or anthropomorphic AI is an “intermediate state” between “mindless” and “mindful”, aiming at manifesting interactive consciousness and accomplishing a variety of cognitive tasks. This division is conducive to the healthy development of the new generation of AI, which not only avoids the “interpretation gap” between two kinds of intelligence due to evolution, but also avoids technical barriers and ethical dilemmas.
点击查看大图
下载:
沪公网安备 31010102003103号