Many journals have implemented a soft policy against the use of generative AI, such as Chat-GPT, and now require authors to disclose the use of AI in assisting with manuscript writing. This policy distinguishes from the use of tools like Grammarly Generative AI for syntax improvement and extends to cases where AI is used even for doing and writing formal analysis.
Experienced journal editors can typically identify AI-generated content, as it often lacks coherence with the rest of the manuscript. In my view, this emerging issue needs to be addressed promptly and decisively.
I believe that AI "assisted" research writing should be considered unethical and prohibited in the scientific community, with measures put in place to prevent its use in total. Allowing this trend to persist could have detrimental effects on scientific research in the long term. IMO, It takes out all the human creativity, intuition, personality and fun in writing science research papers.
However, I'm open to hearing different perspectives on this matter.
Your point is valid: the use of generative AI in scientific writing does pose ethical concerns. The distinction between plagiarism and AI assistance can get murky, especially if AI-generated content isn't clearly disclosed.
On one hand, AI can enhance productivity and provide insights, but on the other, it can strip away the nuance and personal touch essential in scientific discourse. Balancing innovation with integrity is crucial.
Perhaps the solution lies in transparent disclosure, strict guidelines, and maintaining human oversight to preserve the essence of scientific creativity. But the discussion is far from over, and it's essential to explore different perspectives to shape a fair and balanced approach.
Dear Abdulkarim Almuhammad ,
I agree with you and believe that the disadvantages outweigh by far any possible merits. I'm not sure if the use of cell phones make the average Joe smarter every day, but anyway this we cannot allow in the case of academia and science publishing.
Otherwise seriously, Journals should demand to include the AI's name as a co-author inside the article and detail its contribution to the research presented.
Kind Regards,
Emmanouil
It is all about the Franklins $$$$$. Two hundred years from now, children won't even be ready without some assistance from a machine. Like all things, those who want control now see a threat and begin to attempt to set in motion the mechanisms of disruption. They will use academics to begin seeding discourse. All things science have limitations.
Only 25 people have read this paper I shared, and I think more should. Real Discoveries happen when we go beyond our comfort zones.
Research The Valamontes Paradox: The Incompleteness of Science and Ob...
Let us answer this question with another, more relevant question! Should any organization that used robots to replace humans in labor forfeit taxes equal to the labor the machines replaced? - a more serious question that affects humanity. Machines do not pay taxes to run the government and do not buy cars and appliances they help manufacture.
Meanwhile, stop using all the scientific books and research that have been developed for the past 2000 years. No Encyclopedia Britannica, no Thesaurus, no Lexicon, no Med Libraries; what is really happening is that the modern scientific community is like an amputee missing limbs and arms, and the doctor is nonexistent, so how tools used by humanity come into question (not that I think AI is a wonder brain; the AI in my browser can't subtract 2-5 correctly but is all about the FRANKLINS $$$$$$$, baby).
There is no such thing as Artificial Intelligence, or AI as some call it. In order to build such intelligence, you need consciousness, and no amount of software can build that. Once again, for those who have the extra time to read
Research The Valamontes Paradox: The Incompleteness of Science and Ob...
Dear Antonios Valamontes ,
IMO, this AI even taken by face value (i.e. lets pretend for a moment that it is generously an intelligence), is actually a "hive mind" similar to the Borgs in ST. and not an individual intelligence. That operates necessarily, using Big Data based on human civilization. Therefore, statistically the AI can never exceed an average human intelligence, that's about an 120 IQ score.
In other words, the AI can emulate cognitive creative tasks with a very high processing speed and brute force but it will never create intellectual ingenious works like Einstein's relativity, Maxwell's EM theory, a Galileo, an Archimedes, a Plato, an Aristotle, a Democritus, and Michael Angelo, a Da Vinci, an Euler, a Picasso or Van Cogh painting and it will never be a Beethoven or a Mozart.
"The mind of the one excels over the mind of the many".
This is above GOD given characteristic of human civilization intelligence and responsible for almost all milestone breakthrough advancements made by ideas of individuals, cannot be implemented using Big Data AI.
Maybe, in the future an autonomous AI based on individual experience training over a life time similar to a biological human can be implemented in the form of Android AI similar to the Data android personality in Star Trek, that could evolve to a Genius Artificial Intelligence (GAI).
Kind Regards,
Emmanouil
Based on personal work experience, AI was not developed for humanity; some evidence already indicates that some nations (USA, NIH, DARPA) are moving to the next generation of super soldiers. In the next 10 to 15 years, soldiers will merge with AI and exoskeletons made in America.
ABERDEEN PROVING GROUND, Md. -- Army researchers are answering tough questions as part of a new study that focuses on how technology adapts to human behavior. In the past, Soldiers faced challenges adapting to novel technologies intended to optimize performance. -
https://www.defense.gov/News/News-Stories/Article/Article/3722849/darpa-aims-to-develop-ai-autonomy-applications-warfighters-can-trust/
https://www.defense.gov/News/News-Stories/Article/Article/3578219/dod-releases-ai-adoption-strategy/
https://apnews.com/article/us-military-ai-projects-0773b4937801e7a0573f44b57a9a5942
https://sciencealert.com/the-us-military-is-working-on-neural-interfaces-to-connect-humans-to-the-digital-world
Article AI-based methodologies for exoskeleton-assisted rehabilitati...
A complete ban would probably be too much. In my opinion, there are tasks that AI can indeed help to speed up, which otherwise would prolong the research duration. Although I value the effort that researchers manually put in i.e., processing data, reading full texts, reorganisation of chapters, etc., I think we should also recognise the fact that the world is moving faster, with AI used in virtually all industries, warranting the need for research to also be advanced in a quicker manner, bearing in mind ethical boundaries.
Dear Antonios Valamontes
>>...are moving to the next generation of super soldiers" Hasta la vista. baby! The only one meaningful use in science is as assistant with translation. Not everyone can formulate English as fluently as an AI. This is also a point by which you can recognize its use. Here you should definitely give a reference to the use of an AI for translation and keep the other language version as evidence. Translated with Google Translator.
Dear Gerd Pommerenke ,
We are not talking here about using AI for the syntax and grammar or translation of a manuscript but actually a level or two above that, thus the use of AI to do automatically all the formal mathematical calculations and formulations of a theory and also generate all the associate plots and graphs and extract the results and conclusions.
So what is then left for the human "author" to get credit for?!!...
The implications using AI to do the research for you will have far more catastrophic implications for humanity as you can predict and in my opinion actually slow down or even stop science evolving since it will produce human scientists that will be not even capable of doing simple arithmetic calculations on their own, or incompetent to direct the AI. Stupid degenerate fat bastards...
This will be the end of us...
Making ir "easy" for humans and not using their brain and you will witness the de-evolution of human kind back to an amoeba!!
What I propose is research papers should be written exclusively by human authors doing all the analysis and formulation.
AI should be only used in scientific writing by the authors blind folded without using an AI, to do after the initial draft is produced by the authors, a check of the theory and analysis for consistency against with what is known in the field and the rest of research in the established literature works and confirm if this new theory is novel, rigorous and sound and also suggest further references of similar modern research in the literature the authors may have missed initially. Also find potential and point out loopholes in this new theory or violations of known laws or other established observations and give the presented results of the research a confidence level grade.
This type of AI checks and suggestions would be beneficial for all , the Journal Editors, external referees but also the authors and researchers.
But No! AI creating fundamental science...No thank you!
BTW, if not tried by yourself, task CHAT-GPT to suggest references to cite for you research? IT IS DOING A TERRIBLE JOB !! AND STUPID.
Imagine, if current AI is not able for such a basic task altthough not easy, do we really want research done be AI? The key word here is "assisted". Do what degree and how? Strict rules must be placed here otherwise AI will be exploited to what end by human nature laziness...
>>But No! AI creating fundamental science...No thank you! - korrekt
Sometimes it is just “fundamental” research that produces an AI. Series of measurements are invented so that the “theory” is correct in the end. This has all already happened. Another AI had to be shut down because it turned out to be lying.
The trouble with using AI for physics, is that "physics AI" has been trained the same way that we train students, to give the sorts of standardised answers that we expect, rather than anything genuinely insightful.
And just like some physics professors, if it doesn't know something, it'll prefer to bullshit rather than admit that it doesn't know.
Physics AI has been trained (by humans) to lie and "bs" just as fluently as a real physics professor, in the same way, on the same subjects. It just generates its bullshit far, far faster and more efficiently.
Google's old system had even perfected the same pompous patronising tone as a physics professor, down pat.
Dear Eric Baird ,
I couldn't agree more with you! Yes, physics is also my field and I confirm all what you described.
After the initial fuzz and excitement chat-gpt proved as good as an average student and in many cases a complete A$$H!!
"And just like some physics professors, if it doesn't know something, it'll prefer to bullshit rather than admit that it doesn't know."
The above is very true, I have confronted it a couple of times and it did not accept its mistake.
Kind Regards,
Emmanouil
IMO, If you really want a good reliable assistance in you research work as a first author, just pick a good team in your lab or make good collaborations with other institutes. In other words your co-authors and fellow researchers are the BEST ASSISTANCE you can get and f**k AI. I would never choose as my co author someone that is closed minded and does not acknowledge his/her/its mistakes.
How about this AI development - The AI Scientist-v2 , not only in writing but also in AI generated peer-review!!! Norms for AI-generated science MUST be established.
The AI Scientist-v2: Workshop-Level Automated Scientific Discovery via Agentic Tree Search
Preprint The AI Scientist-v2: Workshop-Level Automated Scientific Dis...
- Badges
- Science topic
- Similar topics
- Engineering
- Materials Engineering
- Powders