Editor-in-chief’s foreword

1. The challenges posed by ChatGPT and the impact of the artificial intelligence (AI) field

Among the wide range of topics covered in the previous volume (Volume 45), a notable number of articles addressed the issues related to various artificial intelligence (AI) learning techniques and classification methods. While that is for the research topics addressed in the published papers, in practice, one of the challenges that we faced in the last year was the potential use of ChatGPT chatbot for writing research papers or parts of a paper by the authors. As a disruptive technology, at some point in time, especially in 2023, ChatGPT was at the forefront of many discussions. Many practitioners in the field opined that the use of ChatGPT might not be regulated properly and there could be issues related to copyright infringement. Eventually, even some plagiarism checker software tools (for instance, Turnitin) today include a score for the predicted portion of an article that is (suspected to be) generated using ChatGPT-like tools. Over the course of time however, our understanding is that when the platform is offered to all and when there are thousands of potential users of the same platform, the facility is basically open for all (whoever likes to use it)! This basically would limit the effectiveness of using a ChatGPT-like platform for writing a portion of similar-topic-based papers. There could be increasingly more similarities in the texts as such tools do not possess real intelligence anyway! Eventually, the authors would be bound to modify the primarily generated text and give it their own tone for presentation, which will make the difference. This then is no more than just another tool added to assist human beings and their real intelligence, without really diminishing that.

There is a camp of over-enthusiasts who would like to present AI-based techniques to be superior or those who predict that it would surpass human intelligence, but the fact of the matter is that there are some goals that may never be achievable via AI. Again, even if something close to human intelligence is achieved, it would be a flawed mechanism. Why? This is because the limitation of AI techniques is obvious.

Take for instance, ChatGPT or anything like ChatGPT that may appear in the coming future. When it first hit the tech domain, people got fascinated seeing a poem ‘written by’ ChatGPT but quickly it became apparent that when hundreds and thousands of users make the same query, the outputs do not get any better! Again, a critical limitation of such a tool is that it does not really understand the meanings of the words like a sane human being would do, i.e. it does not have the cognitive capability or consciousness of a human being. It is quite impossible to programmatically code the context-wise sense of understanding or internal meanings of words for a machine or tool that is built by humans. This constraint will remain because, after all, such a tool’s ‘correctness of output’ is to be decided by human beings, i.e. with their feedback. It is also possible to manipulate the output (for future iterations or use) of such a system by intentional incorrect reward and punishment method, i.e. intentional incorrect feedback by collusion of human users. The real intelligence of a human being does not work in the way as these learning-based or training-based tools work. Even for the learning or training process, the input givers are the human beings, who prepare the training data, which can vary depending on the type or choice of the input givers! Hence, this gap cannot be ever overcome as the ultimate controllers of the tool are humans, both for the inputs and outputs.

Let us take for instance a research work that talks about image classification based on AI techniques to identify criminals and fugitives in a crowd. Such a work may talk about installing cameras in airports, rail stations, streets, and various locations to get live videos and images of crowds which would then be assessed by the AI-powered mechanism/tool. Whatever efficiency is shown for any such article that we may have published or other journals have published, while academically this may have some value, in reality, fully relying on such techniques can be seriously disastrous. Why? Because any such work would show at least a percentage of false positives. The accuracy of any such technique cannot be hundred percent and hence, even if there is a single human being who is incorrectly identified (from an image) as a criminal, it can have a tragic consequence (for an innocent human being’s life). Hence, human involvement with human intelligence would be needed on top of the AI-based intelligence technique to achieve accuracy. AI-based tools could simply be a supporting mechanism in such a field. As a research paper, we may appreciate the results but in real life, it may not be a usable or fully reliable tool or technique.

Let us take another case. It has been reported very recently that AI-based mechanisms have been used for target acquisition in some war situations [1]. Any military would need proper and accurate intelligence to hit a target but when an AI-based mechanism generates thousands of targets simply based on an input model or prediction algorithm, it may not translate to the accurate or valid military target but rather could cause maximum chaos, i.e. no better than a random target selection mechanism without accurate intelligence from the ground. Even with billions of dollars in spending, if human beings are not in the loop, AI-based techniques can cause more harm to human beings than good, i.e. what is not good would outweigh what is good (whatever way ‘good’ is defined in this context).

2. The takeaway

AI-based mechanisms have enormous potential indeed. Nobody argues against that; however, the application and implementation of AI-based techniques must be supervised by human’s real intelligence, which is superior. Again, human intelligence can also be influenced by one’s moral compass, which would regulate various decision-making processes.

We hope that the journal’s future contents will also be useful for our research community. Finally, we really appreciate the authors who contributed to this journal with their quality works, the reviewers whose insightful opinions are highly valued in our decision-making processes, and the journal staff who work relentlessly to support the smooth functioning of the journal.

Additional information

Notes on contributors

Al-Sakib Khan Pathan

Al-Sakib Khan Pathan is a Professor at CSE department, United International University (UIU), Bangladesh. He received PhD in Computer Engineering in 2009 from Kyung Hee University, South Korea and B.Sc. in Computer Science and Information Technology from Islamic University of Technology (IUT), Bangladesh in 2003. He has served as a General Chair, Organizing Committee member, and Technical Program Committee member in numerous top-ranked international conferences/workshops like INFOCOM, GLOBECOM, ICC, etc. He was awarded the IEEE Outstanding Leadership Award for his role in IEEE GreenCom’13 and IEEE Outstanding Service Award in IEEE 21st IRI 2020. So far, he has delivered over 30 Keynotes/Invited speeches at various international events. He is currently serving as the Editor-in-Chief of International Journal of Computers and Applications, T&F; Associate Editor of Connection Science, T&F; Editor of Ad Hoc and Sensor Wireless Networks, Old City Publishing, and International Journal of Sensor Networks, Inderscience; Guest Editor of many top-indexed journal special issues, and Editor/Author of 34 books. One of his books was included twice in Intel Corporation’s Recommended Reading List for Developers, 2nd half 2013 and 1st half of 2014; three books were included in IEEE ComSoc’s Best Readings in Communications and Information Systems Security, 2013, several other books were indexed with all titles (chapters) in Elsevier’s acclaimed citation database, Scopus and in Web of Science (WoS), Book Citation Index, Clarivate Analytics, at least one has been approved as textbook at NJCU, USA in 2020, one is among the Top Used resources on SpringerLink in 2020 for UN’s Sustainable Development Goal 7 (SDG7) – Affordable and Clean Energy and one book has been translated to simplified Chinese language from English version. His name appeared on the List of Top 2% Scientists of the World published by Stanford University, USA in 2020 and 2021. He is a Senior Member of IEEE, USA.