摘要
在时间的三种综合方式中,AI没有“未来”这一维度。AI的数据库是过去维度的记忆,它的算法是现在维度的习惯,但它没有未来维度的期待或展望,故AI创作必然要与人进行耦合。人工智能生成影像是装置产生的“技术影像”在软件社会的新形式,扩散模型模拟了技术影像的逆熵过程,生成对抗网络取代了人与装置的互动过程,自动化程序削弱了人(或“凝想者”)的创作功能,影像创作沦为影像计算。计算生成的影像没有“刺点”,也并非“它曾存在过”,而是一种伪造的“似曾相识”。在摄影机成为“自来水笔”、影像如同文字一般灵活写作的半个多世纪后,人工智能生成影像逆转了“影像→语言”的过程,通过提示词生成的影像是一种“语言→影像”,这摈除了社会、情感、价值判断等主观性因素,实现了一种“零度写作”。AI创作不只是一种“自动化美学”,其视觉上的“振动”与审美过程的“情动”都与点彩画派的美学追求如出一辙。这种点彩美学既是一种观察者的技术,主张“看见”影像,而非“看穿”影像;又是一种注意力生产,维持一种动态的、临时的、流动的主体视觉;还是一种打破视觉中心主义的身体实践,践行一种不稳定的身体感受力或具身的身体情动。
In the three ways of synthesizing time,AI does not have the“future”dimension of time,the database of AI is the memory of the past dimension,and the algorithm is the habit of the present dimension.However,it lacks the expectation or prospect of the future dimension,so AI creation needs to be coupled with humans.Artificial intelligence generation image is a new form of“technical image”produced by apparatus in the software society,diffusion model simulates the inverse entropy process of technical image,generative adversarial networks replace the interaction process between humans and apparatus,automated programs weaken the creative function of human(or“einbildner”),image creation is reduced to image calculation.The computed image does not have a“punctum”or“interfuit”,but rather a fake“déjàvu”.More than half a century after the camera became a“Caméra-stylo”and the image became flexible writing-like text,the artificial intelligence generation image has reversed the process of“image→language”,the image generated by the prompt word is a process of“language→image”,which eliminates the subjective factors such as social,emotional and value judgment,and realizes“zero-degree writing”.AI creation is not only a kind of“automation aesthetics”,its visual“vibration”and the“effect”of the aesthetic process are the same as the aesthetic pursuit of pointillism.Pointillism aesthetics is not only an observer’s technology,advocating“look at”the image rather than“look through”the image;but also a kind of attention production,maintaining a dynamic,temporary,flowing subject vision;and also a kind of body practice that breaks the visual centralism and practices an unstable body sensibility or embodied body affact.
作者
范虹
FAN Hong(School of Journalism and Communication,Chongqing University,Chongqing,401331)
出处
《深圳大学学报(人文社会科学版)》
CSSCI
北大核心
2024年第3期152-160,共9页
Journal of Shenzhen University:Humanities & Social Sciences
基金
教育部人文社会科学研究青年基金项目“数字媒体时代互动影音的数据库叙事研究”(21YJC760016)。
关键词
人工智能生成影像
技术影像
零度写作
自动化美学
点彩美学
artificial intelligence generated image
technical image
zero-degree writing
automation aesthetics
pointillism aesthetics