On September 10, the Chinese Academy of Sciences issued an integrity reminder, focusing on the possible problems and challenges in the application of artificial intelligence technology in the whole process of scientific research activities. Regulate the use of artificial intelligence in many aspects, such as topic selection and research, data collection, icon making, achievement writing, peer review, etc.
Recently, the "AI Scientist" jointly developed by Sakana AI Company of Japan and scientists from Canada and United Kingdom has appeared and attracted the attention of the scientific community!
Source: Screenshot of Xinhuanet
It can independently complete the whole process from reading the literature, proposing hypotheses to experiments, proposing solutions, and writing papers, and can generate 10 papers in one go!
Some people exclaim that the era of scientific research automation is coming; Some people also ridicule that this is just an advanced version of the "AI essay generator", and there is still a big gap between it and scientists!
Although this "AI scientist" is not perfect, the potential is huge. This is no doubt a testament to the growing application and impact of artificial intelligence (AI) technology in the field of scientific research.
It is not only a "magical" tool, but also a double-edged sword.
AI has become a power for scientific research
Artificial intelligence not only has powerful data processing capabilities, but also has unique advantages in model optimization, performing repetitive tasks, discovering hidden patterns, and even predicting the results of scientific experiments.
In fact, AI is already playing a practical and effective role in the field of science.
In the face of huge literature resources, retrieval and screening often consume a lot of time for researchers. However, with the help of AI, it is possible to quickly query, read, and even analyze massive amounts of literature, which greatly improves the efficiency of literature reading.
United Kingdom-based DeepMind has developed "AlphaFold", which has predicted the structure of more than 200 million proteins in about 1 million species, covering almost every protein that has been cataloged by the scientific community!
Source: Screenshot of CCTV news
It can be said that AI can help researchers complete a large number of mechanical, arduous and repetitive tasks to a certain extent, allowing them to save more time to think about more complex and high-level scientific problems.
Not only that, but AI can also span multiple different disciplines and accelerate the cross-integration of different disciplines. With in-depth learning, powerful algorithms can even change the traditional research path and open up an intelligent and automated research path.
But while providing more convenience for scientific research, AI has also raised many problems.
The academic crisis brought about by AI
With the reduction of the cost of AI-generated papers, some people use AI technology to fabricate experimental data, create academic garbage, and produce a large number of fake research papers.
In April this year, a reporter went undercover to the paper writer group, and the writer made it clear that since the original author did not provide data, let the AI compile some data, and then let the AI analyze the data according to the compilation.
Source: Beijing News
Previously, a Harvard Kennedy School's "Misinformation Review" study found that AI-generated fake research papers were infiltrating academic search engines such as Google Scholar. These fake research papers not only lead to wasted resources, but also undermine public trust in science.
These low-quality or even worthless papers may exacerbate the formation of "academic bubbles", leading to lower academic standards and hindering the progress of science to a certain extent.
And, most of the time, due to the lack of a proper understanding of scientific concepts and context, there is a phenomenon of "serious nonsense".
For example, the AI mouse picture that caused a stir in the academic community some time ago not only seriously inconsistent with scientific facts, but also had a large number of English spelling errors, and was quickly retracted after only 3 days of publication.
(PS: Let's find fault together, the picture is wrong at the end of the article)
Source: References[2]
In this regard, Nick · Bostrom, head of the Institute for the Future of Humanity at the University of Oxford, believes that if we introduce superintelligence before solving the control problem, the future development of artificial intelligence research will pose a major threat to human survival.
Multi-party regulation of AI use
With the wide application of artificial intelligence in the field of scientific research, relevant departments and scientific fields have paid more and more attention to its standardization.
On July 13, 2023, the Cyberspace Administration of China, together with the National Development and Reform Commission, the Ministry of Education, the Ministry of Science and Technology, the Ministry of Industry and Information Technology, the Ministry of Public Security, and the State Administration of Radio, Film and Television, announced the Interim Measures for the Administration of Generative AI Services.
On August 28, 2023, the People's Republic of China Academic Degrees Law (Draft) clearly emphasized that the use of artificial intelligence to write dissertations is academic misconduct.
In December 2023, the Department of Supervision of the Ministry of Science and Technology issued the Guidelines on Code of Conduct for Responsible Research (2023) (hereinafter referred to as the "Guidelines"), which added restrictions on the use of "generative AI" in research behavior.
In August this year, the European Union's Artificial Intelligence Act came into effect, which is the world's first regulation to comprehensively regulate artificial intelligence.
Despite this, the establishment of regulations at this stage still cannot keep up with the rapid development of artificial intelligence. This requires researchers to have certain ethical norms and always put academic responsibility and academic conscience in the first place, so as not to be led astray by disruptive technologies such as artificial intelligence.
Research is not only a series of tasks to complete independently and efficiently, but also a complex and innovative undertaking. At present, AI is still at the forefront of exploration, and scientific research and innovation still need people to lead the way, analyze, think and make decisions.
Zhuangzi said: Things are not things in things.
As a human invention, AI is not good or bad, and it is the right solution to use it for me.
Inspiration and innovation are always the patents of human beings, and it is the dignity that every researcher should maintain to maintain a strong thinking power in the face of artificial intelligence.
(Have you figured it out?) ）
Resources:
[1] Xu Dongbo. The Logic, Risks and Governance of AI-driven Scientific Research[J]. China Science and Technology Forum, 2024(5): 120-129.
[2]https://www.frontiersin.org/journals/cell-and-developmental-biology/articles/10.3389/fcell.2024.1386861/full
Sources: Academic Bridge Comprehensive Zizhihu, CCTV, The Lancet, Beijing News, Qubit, Lilac Garden