Text|A Zhi
Editor|Wang Xiaokun
In the field of advertising and marketing, the current copywriting ability of artificial intelligence is equivalent to 2.47 years of human experience. Although humans are still stronger than AI in copywriting creativity and insight, the latest experimental results show that when it comes to copywriter identity, industry experts/workers are able to identify AI and humans, and consumers can no longer distinguish between them.
The above conclusions come from the latest paper published by Chinese Minmin University, Ningbo University and Miaozhen School of Marketing Science in Business Economics and Management, "A Comparative Study of Artificial Intelligence and Human Creativity: Based on the Dual Perspectives of Experts and Consumers" (hereinafter referred to as "Research").
The "Research" focuses on the field of advertising and marketing, setting up a total of 10 advertising copywriting tasks, covering 7 different industry directions, each task contains the creative results of both human copywriting and artificial intelligence copywriting, and adopts the method of double-blind experiment to recruit experts (60 people, including marketing practitioners and university marketing teachers) and consumers (1707 people) to evaluate the difference between artificial intelligence and human copywriting ability.
Example of task prompts input to artificial intelligence, source paper
Artificial intelligence has replaced primary advertising copywriting, and there is no significant difference in the test results of large model applications at home and abroad
Specifically, look at the results of the study.
In terms of copywriting resolution, experts have certain distinguishing ability, and can correctly distinguish between human copywriting and artificial intelligence copywriting in 59% of cases; But consumers can't tell the identity of the creator, and AI-written copywriting is mistaken for human-written (definitely human + probably human) 55% of the time, while only 53% of human copywriting is considered human-written. That is, the copywriting created by artificial intelligence has reached a state that consumers cannot distinguish, in a sense, "passed the Turing test of copywriting".
Image source paper
From the perspective of the working years of copywriting matching, the expert evaluation results show that the corresponding experience of the copywriting generated by artificial intelligence is 2.47 years (standard deviation 1. 58 years), while the corresponding years of experience for human copywriting from college student entries was 3.83 years (standard deviation 2. 19 years). That is, the working years corresponding to the current human copywriting ability are still higher than those of artificial intelligence.
Image source paper
Experts evaluate copywriting professionalism and conduct ANOVA analysis on the results in five dimensions: clarity, completeness, ease of understanding, creativity, consumer insight and business insight. The conclusion shows that on the 10 advertising copywriting tasks, the professional ability score of human copywriting is still higher than that of artificial intelligence, but specific to the dimensions of professional ability, the scores of artificial intelligence copywriting in "clear and complete" and "easy to understand" are not significantly different from human copywriting, and even ChatGPT and Wen Xin Yiyan have higher ability evaluation than human copywriting in the latter, but they are still significantly lower than humans in the three dimensions of "creativity", "consumer insight" and "business insight".
Image source paper
Different from the expert evaluation dimension, consumers cannot distinguish the authorship of copywriting, and the analysis is mainly based on the perception evaluation dimension. The conclusion shows that human copywriting and two types of AI copywriting score extremely close in terms of consumer perception level. Specific to the comparison of various sub-dimensions of consumer perception, human copywriting only performs better on a few tasks, most tasks have no significant perception difference, and even score low in "easy to understand", "relevant", and "willingness to share", and consumers have better perception of artificial intelligence copywriting.
Image source paper
In addition, it is worth noting that the 20 artificial intelligence creation texts in the "Research" are from ChatGPT and Wen Xin Yiyan (10 each), but in the above expert and consumer evaluation and analysis conclusions, ChatGPT and Wen Xin Yiyan did not show significant differences in professional ability. This means that, at least in the field of advertising copywriting, the current domestic and foreign large model applications are similar in this professional level.
The marketing industry is facing "technical unemployment", and human-intellectual collaboration will create a new batch of job demand
The daily work of the advertising and marketing industry revolves around the purpose of communication, which has a natural "openness", which can be defined as "exposed" work activities, which are easier to be pre-trained and absorbed by large models. This means that in the future, work activities and occupational content "exposed" to artificial intelligence will be easier to replace, and shocks have already begun to appear.
Although there are advantages in professional dimensions such as creativity and insight, and the evaluation of years of experience is higher than that of artificial intelligence, the "Research" team also pointed out that "considering that the speed of improvement of human professional ability may be much lower than that of artificial intelligence, perhaps the leading edge of human beings in this ability will disappear in the near future." And with the "emergence" and "prompt words" of artificial intelligence, the creative output ability of artificial intelligence will be further improved.
According to a survey of 1,000 U.S. business leaders by third-party platform Pollfish, 49% of businesses are currently using ChatGPT, while 25% have saved more than 7% by using ChatGPT. $50,000.
Whether it is ChatGTP or Wenxin, the landing and iteration speed of large models and applications in various fields at home and abroad are changing with each passing day, standing at the node of marketing industry change, the development of generative artificial intelligence will bring a new round of human capital replacement almost become an industrial consensus. The research team believes that due to the high efficiency of artificial intelligence and the uninterrupted rhythm of 24-hour work, the "technical unemployment" brought by artificial intelligence to the marketing industry will dominate, but at the same time, new jobs will be created, such as chief artificial intelligence officers (CAIO), AIGC work processors, AI trainers, etc., and human-intelligence collaboration will become the mainstream of the marketing industry.