ChatGPT shortly after the appearance，once became a sensational event in the field of culture and technology，It has caused widespread discussion in society。This system can automatically create text based on written prompts，Welcome to many users。Domestic and foreign technology companies also strive to follow suit，It has launched similar artificial intelligence systems。but，Genomic artificial intelligence system like ChatGPT，There are many problems at the social level and academic circles。So，How will this technology affect the role of papers in education? Has the performance of the amazing ChatGPT achieve general artificial intelligence?。other，Put aside these more magnificent problems，We also need to focus on the issue of scientific research and integrity caused by ChatGPT。

bet365 best casino games  

In the past few years，Large language models have triggered an increasingly complex academic debate on its ethical impact。With OpenAI released ChatGPT at the end of 2022，This discussion has now become a trend。When ChatGPT receives instructions，You can quickly generate text，and the text is smooth in most languages、turning natural，This expression is impressive。But，In further dialogue，will soon find some of its shortcomings。For example，It may make low -level fact errors，and the same phrase is repeated。Despite this，But when the first chat robot interacts with such chat robots，It will still make it difficult for people to suppress the feeling of admiration。ChatGPT shortly after release，Many people have raised various problems bet365 live casino games on its influence。For example，As these artificial intelligence systems are becoming more and more complete in writing code，Software engineers will unemployment? What impact will it affect knowledge workers?，Some are old -fashioned talks，and some are quite new。

In terms of text generation，One way to look at ChatGPT is，When the academic circles are studying and writing papers，A new type of hand -used tools in the growing tools in the toolbox。For example，It can be regarded as a search engine that can directly answer questions。Some people may use ChatGPT to perform some kind of "creation": provide the first draft of new works。but，ChatGPT also has great limitations。It sometimes gives the wrong answer，It may be too sensitive to the subtle difference in the prompt word，It may also give answers that lack correlation，and given the necessary long story and so on。These defects limited its utility as a research tool。Because of worrying that ChatGPT and similar software may fabricate data in the generated articles，Currently thousands of scientific journals publishers have banned or restricted the submitters using artificial intelligence -driven chat robots。and，Some colleges and universities will also be judged by the operation generated by ChatGPT as plagiarism or plagiarism。

Another way to think about ChatGPT emphasizes its revolutionary。It is true，ChatGPT may not be better than general scientists in writing scientific papers。but but，Some people think，In view of the recent index levels of artificial intelligence，Preparing for the future large language model for the future may be wise。We can see similar development from the chess project。Initial，Only a few people believe that computers can be better bet365 best casino games than human players。But，In 1997，IBM's deep blue defeats the international chess world champion Casparov。Same，In 2016，The alphaGo of DeepMind defeated the Go world champion Li Shishi。So，From the perspective of some technical optimists，Although ChatGPT is currently not perfect，But the direction of the future technological development。They even think，Perhaps writing a scientific papers is just a smart activity that computers may learn better than human performance。

　General artificial intelligence says early 

According to the expectations of technical optimism，ChatGPT is leading to the Kangzhuang Avenue, which is generally universal artificial intelligence? Turing put forward the idea of ​​later called "Turing Test" in 1950，Whether to pass this test has become an important indicator for measuring the intelligence of artificial intelligence。If you want to pass test，Machines need to be able to dialogue with a human dialogue that cannot be distinguished from humans。The difficulty of the project is reflected in，To pass the test，Need to meet two conditions: make appropriate reactions to the differences in human conversations; show a similar human personality and intention。

In the 1970s and 1980s，The study of artificial intelligence in the academic community was full of enthusiasm，But because of the real useful artificial intelligence application system, it was not available，This kind of enthusiasm gradually fades in long setbacks。Perhaps，History always has some cycle similarity。Current，We have experienced an excitement again -deep neural network、New progress of machine learning technology and various artificial intelligence algorithms represented by ChatGPT，once again inspired the public's enthusiasm。But we should bet365 best casino games not be blind and optimistic。Under the incentives of these new progress，There are also some gaps that we can't pass temporarily，Especially some problems in the field of natural language treatment。

The complex language system is a significant feature of human intelligence，And the language and thought have a close relationship，It can even be said to some extent，"Language is the carrier of thought"。If the machine cannot "understand" human language，High -level intelligence cannot talk about，Not to mention GM。First，We think，Due to mathematical reasons，It is still difficult to give a machine writing program that can master extremely complicated and changing human dialogue mode。Next，Human conversation is mixed with personality and motivation，But we still don’t know how to create a machine with human personality and intention。Since the Turing machine cannot master the human conversation behavior，We can get a security conclusion: The current chat dialogue system that has been released cannot reach the so -called general artificial intelligence。Human cognition itself is not just a question about storage and calculation。

，"Chinese House Analysis" of John Searle indicates，Even if the programming method of a machine can perfectly simulate the behavior of human beings in dialogue，Can't really have human intelligence。Sel thinks，The perspective of the understanding of the machine and the human consciousness is doubtful。Although people do a lot of work，But so far, there is no computer system that can achieve simulation simulation that can be close to human dialogue。Even ChatGPT，It is still revealing the fundamental defects of chat robots: more mechanical，Insufficient flexibility。Chatgpt is bet365 best casino games not understanding the meaning of the text and made a correct or even accurate answer with this。As Openai warned: "The answer you get may sound like not or even authoritative，But they are likely to be completely wrong。”

Needless doubt，Today's storage capacity and computing power owned by smart machines have far exceeded the "learning machine" estimated by Tu Ling，But in some situations, there is still not much change。We think，Excessive optimism in this field，On the one hand, it is because of an inconsistent view of human cognitive behavior，On the other hand, I saw strengthening learning、Data set expansion、A series of amazing success obtained in the algorithm update，Progress in the fields of Go or video games。Contrary to this optimism，People have repeatedly failed when trying to make efficient dialogue with humans and even reach the general artificial intelligence stage。Current，Technical Optimism School advocates that GM artificial intelligence is too early。

　Following the issue of scientific research integrity 

Due to the extreme complexity of human language，Various conversation robots are difficult to truly realize natural language processing，Not to be conscious、Emotional、Sensory, etc.。Chatgpt is just a pre -training language model，It can only extract from existing training samples、Integrated information，Instead of really producing new knowledge。On the way to the way，Universal artificial intelligence and super artificial intelligence are out of date。Besides，Chatgpt is not a product that was born，The GPT-3 before iterative has caused a sensation in public opinion。Whenever there are new technologies or products that have been packaged，There will always be a whirlwind，But often Bet365 app download stop at a stop soon。We don’t have to worry about predicting whether a certain technology or product can bring so -called epoch -making significance。

aside the great narrative，Some of the specific consequences behind this issue may be more worthy of our in -depth inspection。Chatgpt will definitely have some positive effects。For example，Education workers can learn to use ChatGPT，Use it as a tool to help students explore a specific problem。But currently the focus is needed，It is a problem of exposed exposure。Generating artificial intelligence has huge potential in text generation and editing，If the user is abused in academic writing or news report，It will bring a lot of adverse effects。So，For large language models and its impact on scientific publication，We need to conduct forward -looking research。The current dilemma is，Before obtaining such research results，We can only make some intuitive on the characteristics and expectations of large language models、Common sense assessment and judgment。

If the generating artificial intelligence such as ChatGPT is used as a research tool，It will cause some concerns in scientific research integrity。For example，This will bring inaccurate facts、Plagiarism、Risks such as fraud and violations of copyright，It may cause science to lose rigor，The labor of science and technology workers may also become un respects。In the field of scientific research，Careful applications for generating artificial intelligence such as ChatGPT，Is a responsible attitude。Many academic journals and scientific research institutions have begun to formulate and implement some feasible solutions，Solution or cope with research integrity related to ChatGPT，But the future prevention and governance work is still long and far away。

　　(This article is the "Research on Ethical Regulations of Human Enhanced Technology" (2023J0083) phased achievement) 

(Author Unit: School of Marxism University in Yunnan University)