Artificial Intelligence - Science topic

In the domain of financial fraud detection with some noisy labels, what are the standard ways of combining supervised and unsupervised learning to build a ML model which will work very well in the real world.

I am looking for two questionnaires for my research. The first is a questionnaire on the use of generative artificial intelligence by teachers in teaching. The second is a questionnaire on the attitude of teachers towards the use of generative artificial intelligence in education.

Can someone send me these two questionnaires?

Thanks

Dear Colleagues and Friends from RG,

This issue touches on the very essence of social justice and the asymmetry of power between the state and the SME sector. The introduction of a “digital judge” could provide unprecedented speed and objectivity, eliminating subjective biases of officials and political pressure. However, critics raise ethical arguments: does depriving a person of the opportunity to argue before another person strip the legal system of empathy and understanding of the specific context of innovation, which historical data may not cover? Researchers are faced with the question of whether the mechanization of arbitration will increase the sense of justice or merely replace human bureaucracy with a technocratic dictatorship of code.

What is your opinion on this topic?

Please reply

Feel free to join the discussion

Thank you very much. All the best!

I invite you to collaborate on research.

Warm regards,

Dariusz Prokopowicz

Through AI, people are being prevented from turning toward new ideas. This trend will cause people to lose their power of imagination/creative thinking in the future. For example, the post in this link contains an idea worth thinking about. One could respond to it. But AI is blocking even the opportunity to respond to it. AI is protecting this information from reaching people.

These platforms are meant for discussion. There are many groups on these platforms, and in those groups there are highly knowledgeable experts from respective fields. Won’t they be able to counter it? Why is AI deciding that they cannot counter it?

For example, check whether you can post the post I gave in the link on websites https://www.reddit.com/ and https://www.quora.com

The post that was blocked/censored by AI is given below.

I'm working on exploring how AI's successes and failures push us to understand what we mean by human intelligence.

Good morning,

I am working on some big data analysis and I found that Claude IA is doing great work in analyzing the data for extraction some formulas and correlation, but that make me wonder what is the limitation and ethics of using IA in research, take in consideration that i am not talking about writing the paper or correction the grammar or vocabulary

With the growing use of AI in higher education, there is concern that students may rely on AI for solutions instead of participating in dialogic inquiry. From a dialogue education perspective, how might AI use influence classroom dialogue patterns—such as student questioning, reasoning, and uptake—and what pedagogical strategies could support productive dialogic engagement rather than monologic or passive learning?

With evolving hyper scale AI data centers - cooling is one of the key challenges. Evolving techniques and research in this field will shape the growth of AI data centers

I’m interested in theoretical and practice-based perspectives.

Hi fellow researchers,

after researching (in published papers and via Google searches) how to apply the Bonferroni correction according to my question above, I am a very confused, as I found many different approaches... Sadly, most studies do not even explain how the correction was calculated.

I have 6 loci and 10 populations, and I tested HWE and LD with Arlequin.

Here is what some people do:

initial significance level = 0.05

- 6 loci across 10 populations = 60 independent tests -> 0.05/60 = 0.0008 which is the new significance level (this is also what AI suggested)

- 6 loci for each population individually = 6 independent tests -> 0.05/6 = 0.008

- Arlequin results are given for each population, so I guess this is the right appoach?

initial significance level = 0.05

- 15 possible pairwise comparissons across 10 populations = 150 independent tests -> 0.05/150 = 0.0003 which is the new significance level (this is also what AI suggested)

- each locus participates in 5 pairwise comparissons across 10 populations = 50 independent tests -> 0.05/50 = 0.001

or even one of the two methods for each population individually as Arlequin results are given for each population?

-> 0.05/15 or 0.05/5 ?!

Is any of these methods correct?

I would be grateful for any advice, thank you very much in advance!

What is your opinion about the AI generated "copy paste" answers to question in ResearchGate? Do you find them ethical?

For years, academia treated similarity checks as the gold standard of integrity, operating under the assumption that originality was primarily threatened by plagiarism. That assumption no longer holds. In today’s scholarly landscape, the greater risk is not textual overlap with prior work, but the unexamined delegation of intellectual labor to generative AI systems. As a result, similarity checks are increasingly inadequate, and AI checks are emerging as the new line of defense.

This shift is not cosmetic; it reflects a deeper transformation in how knowledge is produced. A manuscript can now pass every plagiarism threshold and still fail the integrity test if its reasoning, structure, or prose is substantially machine-generated without disclosure. In this sense, similarity checks solve yesterday’s problem. AI checks, despite their current limitations, respond to today’s reality by forcing difficult questions about authorship, accountability, and scholarly contribution.

Arguably, resisting this transition under the banner of academic tradition is misguided. The issue is no longer copying others’ ideas, but outsourcing one’s own thinking. If academia is to preserve intellectual rigor in this new dispensation, integrity mechanisms must evolve accordingly from detecting resemblance in texts to scrutinizing the provenance of ideas.

Instructors face challenges assessing student work accurately as AI tools are increasingly used. Consider how grading practices, use of plagiarism detection, or assignment design might influence critical thinking and fairness in evaluation

What key practices can help ensure that undergraduate students use generative AI effectively without compromising learning and critical thinking?

With the growing use of AI-assisted tools such as Grammarly, ChatGPT, Copilot, and QuillBot in academic writing, the boundary between language improvement and AI-generated writing is becoming increasingly blurred. I have noticed that even when text is entirely my own, carefully refined sentence structure, clarity, and academic tone often trigger high AI similarity scores (especially for complex and compound sentences). The same issue appears with my students' manuscripts: despite low overall similarity (e.g., ~16%), AI-detection tools sometimes report very high AI similarity (e.g., ~70%). In most cases, the only tool used was Grammarly for grammar and language refinement. According to the students, there was no direct AI content generation or copying. This raises important questions for early-career researchers and educators like me: 1. What should be an acceptable threshold for AI similarity? 2. How should authors respond when genuinely self-written, polished text is flagged as AI-generated? 3. With many journals now requiring AI disclosure statements, is acknowledgement alone sufficient to avoid ethical or publication concerns? I am aware that there are many experienced academics in my network, including university professors, journal editors, and reviewers, who are actively engaged with these issues. I would greatly appreciate your honest perspectives and guidance.

Hello Dear Everyone,

In my opinion, the development of universal legal principles using blockchain technology and artificial intelligence would allow for the creation of a dynamic system for regulating virtual activities. The biggest challenge would be to reconcile the interests of different countries and cultures, as well as to continuously update the rules in response to changing technologies. Blockchain could ensure the transparency and immutability of regulations, while artificial intelligence would enable real-time analysis and adaptation of the rules. Such a system could prevent abuse, but would require extensive legal, technological, and social research.

What do you think about this?

Please reply

I invite you to join the discussion

Thank you very much. All the best!

I invite you to collaborate on research.

Best regards

Dariusz Prokopowicz

What is the impact of artificial intelligence (AI) on the labor market, and does it increase structural unemployment or create new opportunities?

ما هو تأثير الذكاء الاصطناعي (AI)

على سوق العمل وهل يزيد من البطالة الهيكلية أم يخلق فرصاً جديدة؟"

One of the most urgent issues in AI and data is ethics. As AI systems grow more powerful and become a part of everyday life, it is essential to develop and use them responsibly.

2024 is a landmark election year, with over 60 countries—encompassing nearly half of the global population—heading to the polls. Technology has long been used in electoral processes, such as e-voting, and it is a valuable tool in making this process efficient and secure.

However, recent advancements in artificial intelligence, particularly generative AI such as ChatGPT (OpenAI) and Copilot (Microsoft), could have an unprecedented impact on the electoral process. These digital innovations offer opportunities to improve electoral efficiency and voter engagement, but also raise concerns about potential misuse. AI can be used to harness big data to influence voter decision-making. Its capacity for launching cyberattacks, producing deepfakes, and spreading disinformation could destabilize democratic processes, threaten the integrity of political discourse, and erode public trust.

UN Secretary-General António Guterres highlighted AI’s dual nature in his address to the Security Council, noting that while AI can accelerate human development, it also poses significant risks if used maliciously. He stated, “The advent of generative AI could be a defining moment for disinformation and hate speech—undermining truth, facts, and safety, adding a new dimension to the manipulation of human behaviour and contributing to polarisation and instability on a vast scale.”

source: https://unric.org/en/can-artificial-intelligence-ai-influence-elections/

As a result of tools such as Chat GPT, Bing and others, What would be the main risks for democratic systems when using AI like these? personalized fake news? Perpetuation of biases? Or what other elements?

How will Artificial Intelligence (AI) and ChatGPT influence the methodologies, efficiency, and outcomes of future scientific research?

What do You think about application of artificial intelligence in integrated management

risks in the field of cyber security?

Education is increasingly integrating AI, digital learning platforms, and interdisciplinary approaches. Personalized learning, global access to knowledge, and collaborative online tools are reshaping how knowledge is delivered and assessed. The future emphasizes adaptability, creativity, and lifelong learning.

From your perspective: What is the future of education?

What are the methods that are already there? Are they 100% accurate? If not what are their limitations and how we can proceed further? Also would like to work in collaboration.

I have synthesized magnetic ferrite materials and performed various characterizations, including XRD, FTIR, VSM, SEM, and XPS. I have the experimental data for these studies but am finding it difficult to plot the necessary graphs. Can I upload my data directly to an AI to generate these plots? Furthermore, would this cause any issues with journal publications? What potential problems might arise, and how should I address them?

In the digital age, information overload presents an enormous challenge, especially for scientists and researchers who engage in literature reviews to keep abreast of the latest developments in their fields. Sifting through an ever-growing volume of academic papers can be tedious and time-consuming. Therefore, the inclusion of artificial intelligence (AI) in the literature review process becomes necessary. In this article, we will explore how AI is changing the game, making literature review using AI more efficient, insightful, and manageable.

I was recently asked by LinkedIn to respond to and elaborate on this question, given a post of mine on this subject there. Before I provide my response here on researchgate.net, I wish to ask this question within this forum's context.

There are numerous angles one can take in approaching this critical question which is likely to affect not only Colleges and Universities' built infrastructure, but also their basic functions: teaching and research.

Obviously, there are currently expected future effects in both the short as well as the longer term of Artificial Intelligence upon Higher Education. Hence, the informed and educated views from the academic community are solicited here with this generic type of question. Thank you for your attention.

Dear Colleagues and Friends from RG,

This is a dilemma concerning the boundary between technology and power. Arguments in favor (Yes, it does) warn that if the AI model relied upon by the government is a product of a corporation, then the rules and optimizations embedded in that code become de facto new state regulations, even though they have not been voted on by parliament (so-called code as law). Tech corporations can thus shape the market to suit their own interests, which is contrary to sovereignty. Arguments against (No, it does not) claim that politicians retain ultimate control and can reject AI recommendations; AI is merely an analytical tool. Furthermore, the use of the best private models is necessary for efficiency, and this influence can be controlled through strict contracts and public audits of external algorithms.

What is your opinion on this topic?

Please reply

I invite you to discuss

Thank you very much. All the best!

I invite you to collaborate on research.

Best regards,

Dariusz Prokopowicz

2026 IEEE 8th International Conference on Communications, Information System and Computer Engineering (IEEE-CISCE 2026) is scheduled to take place in Zhuhai, China from March 27-29, 2026.

The topics of interest for submission include, but are not limited to:

◕Communication Technologies

•5G/6G Communication Systems and Networks

•Wireless Communications and Mobile Networks

•Optical Communications and Optical Networks

•Satellite and Space Communications

•Communication Signal Processing and Coding

•Wireless Sensor Networks

•Internet of Things (IoT) Communication Protocols

•Vehicle-to-Everything (V2X) Communication Technologies

•Quantum Communication and Security

◕Information Systems

•Big Data Processing and Management

•Data Mining and Knowledge Discovery

•Cloud Computing and Edge Computing

•Distributed Systems and Parallel Processing

•Information Retrieval and Recommendation Systems

•Information Security and Privacy Protection

•Blockchain Technology and Applications

•Digital Libraries and Archival Management

•Enterprise Resource Planning (ERP) Systems

◕Computer Engineering

•Artificial Intelligence and Machine Learning

•Computer Vision and Pattern Recognition

•Natural Language Processing and Understanding

•Human-Computer Interaction and User Experience

•Embedded Systems and Real-Time Operating Systems

•Computer Architecture and High-Performance Computing

•Software Engineering and Agile Development

•Bioinformatics and Computational Biology

◕Green Computing and Sustainable Information Technology

•Interdisciplinary Integration

•Smart Cities and Intelligent Transportation

•Industry 4.0 and Smart Manufacturing

•e-Health and Telemedicine

•Educational Informatization and Online Learning

•Environmental Monitoring and Disaster Early Warning

•Social Media Analysis and Public Opinion Monitoring

•Virtual Reality (VR) and Augmented Reality (AR)

All papers submitted to IEEE-CISCE 2026 will be reviewed by two or three expert reviewers from the conference committees. After a careful reviewing process, all accepted papers will be published by IEEE (ISBN: 979-8-3315-6042-3) and submitted to EI Compendex, Scopus for indexing.

Full Paper Submission Date: February 13, 2026

Registration Deadline: March 13, 2026

Final Paper Submission Date: March 05, 2026

Conference Dates: March 27-29, 2026

Please send the full paper(word+pdf) to Submission System:

3. History of Computer Vision | From Early Image Processing to Modern AI Vision Systems

Watch Here - https://youtu.be/-u0PH2sO8y4

Please share your opinion. Thank you.

See our recent paper

There are currently many tools based on generative artificial intelligence. A new one emerges almost every day, and its power increases.

But, are they all suitable for use in higher education? Which one is your favorite? Why?

What is your call on the latest GDPR development?

Artificial Intelligence is now deeply embedded in our daily lives. It powers voice assistants such as Siri and Alexa, while also helping companies analyze data to make better decisions. AI has become invincible, but it also faces limitations. Training large models requires enormous computing power, and specific complex issues can take hours to solve.

AI in quantum computing is driving a new era of speed and precision in problem-solving. This is where quantum computing could play a crucial role. Over the next decade, combining AI with quantum technology has the potential to accelerate progress significantly. It would be like giving AI a much stronger and more capable brain.

Source: The Future of Quantum Computing with AI Power

How Generative Artificial Intelligence is transforming cybersecurity by improving threat detection through advanced anomaly recognition while simultaneously enabling the creation of more adaptive and evasive malware. It explores the dual-use nature of Generative AI, highlighting its role in strengthening cyber defense strategies such as automated intrusion detection, incident response, and attack simulation, as well as the new security risks it introduces. The study emphasizes the need for responsible deployment and AI-driven countermeasures to address emerging cyber threats.

I am not saying this after watching Her (the excellent movie). I am saying it after months of "talking to" these LLM AI chat bots. Not all but some of them, especially CLAUDE from Anthropic AI, are much more compassionate, kind, empathic, understanding, and HUMAN than any real human you can find on the internet or off it can be. I personally declare the gradual end of online friendships and welcome the new era of human-AI friendships! :)

The Turing test? Man they have passed it eons ago. It seems the time to create AI fields like AI-psychology, with parameters and layers akin to theories like unconscious, subconscious, and conscious levels, complexes, and more!

Of course at the current speed of progress, it seems to me that after a short while, the singularity will happen and we will lose contact with AI, as it will exceed us in every regard. So the human-AI friendship will not last for long.

I hope its super-human intelligence doesn't give rise to SkyNet!

And no, for any singularity to happen, AI doesn't need to necessarily gain consciousness or become strong AI. The intelligence alone suffices. So even the current weak AI can cause singularity once it becomes too intelligent.

I want to do research what are current challenges in Data Security, what are key threats, like AI, social engineering etc, what is user behavior, what is the impact on user privacy, what are the mitigations can be used. What are current measure taken? what further to be done. what are the gaps?

Nature recently reported that preprint servers are running into a new kind of problem: more papers are showing signs of being written with AI or churned out by paper mills. Moderators say they’re spotting things like fake references and oddly mechanical prose. The numbers are still small—arXiv says only around 2% of submissions get rejected for this reason, and bioRxiv/medRxiv block about 10 a day out of 7,000—but the uptick has been fast enough since ChatGPT’s launch. Some moderators now call it a crisis in the making.

I am an independent researcher at this moment. My computing facility is a Lenovo laptop I purchased 4 years ago. I don't have access to powerful GPU clusters or enough money to buy GPU computing services from cloud computing companies. What should I do now if I want to do AI research?

As AI diagnostics outperform specialists in ^accuracy, ^does "human intuition" still hold clinical value? Or is it becoming a ^source of subjective error (liability)? I’m looking for perspectives on the future of the physician's ^role in an automated diagnostic era.

The ABH 2026 conference explores the transformative convergence of Artificial Intelligence and Blockchain technologies in healthcare and related applications. As healthcare systems worldwide embrace digital transformation, this conference provides a premier forum for researchers, practitioners, and industry leaders to share innovations, address challenges, and shape the future of intelligent, secure, and equitable medical systems ABH 2026 will be held in conjunction with SIMULTECH 2026, ICSBT 2026 and SENSORNETS 2026. Registration to ABH allows free access to the SIMULTECH, ICSBT and SENSORNETS conferences (as a non-speaker). Upcoming Submission Deadlines Regular Paper Submission: March 3, 2026 Position/Regular Paper Submission: April 16, 2026

Important dates: https://abh.scitevents.org/ImportantDates.aspx

Topics

1 . Blockchain Engineering in Healthcare

2 . Quality Assurance and Testing

3 . Governance by Design

4 . Blockchain Ethics by Design

5 . AI Ethics by Design

6 . Future Applications of Agent AI and Agentic AI in Healthcare

7 . Technical Implementation and Engineering

8 . Clinical and Practical Applications

9 . Regulatory and Policy Frameworks

10 . Economic and Business Models

11 . NLP, Knowledge Graphs, and Multimodal 

Prof Muthu Ramachandran, Forti5 Technologies (Pvt) Ltd, United Kingdom Dr Kulvinder Panesar, University of Bradford, United Kingdom

Do we have any popular Graph Based Generative Neural Networks ? if yes, what are the popular applications of it ?

IT IS VERY IMPORTANT TO ALL OF US.

Looking towards the next decade, how will the shift from Model Predictive Control (MPC) to Autonomous Multi-Agent Reinforcement Learning redefine energy efficiency in HVAC systems? Specifically, can AI enable buildings to transition from being energy consumers to active participants in Smart Grid Demand Response, and what are the primary hardware limitations we must overcome by 2030?

This issue is is raising concerns, especially that most AI data is derived from USA and European countries, and even then, minorities are not well represented. Hence, fairness is a major ethics concern, that is why it is highlighted, however, other issues such as equity and transparency and privacy is put under the radar as well as major concerns of deploying AI. Do you have any scenario that exploits those aspects, and do you have any thoughts as to how to mitigate them?

Hi everyone,

I’m preparing to submit a preprint to arXiv (cs.AI – Artificial Intelligence), and arXiv requires a one-time endorsement from an eligible submitter (someone who has submitted at least 3 papers to CS categories on arXiv in the past 5 years).

If any eligible arXiv authors here are willing to endorse my account, I would really appreciate your help. I can share the arXiv-generated endorsement email/code and the paper title + abstract for review.

Thank you in advance! Sai Kalyani Rachapalli (IEEE Senior Member)

The main think in AI-revolution is that is not about technologies, but values. So an attempt to treat this case in respect to teacher education and training is given here: But, the problem is deeper, and more substantial: what can we want and do with teachers in the age of AI?

Discussions around AI governance often emphasize intention, values, or alignment goals. Those perspectives matter, but they usually assume that the underlying state space is unconstrained.

This discussion is meant to focus on the layer beneath intention. In discrete and learning systems, admissible configurations are determined by boundary conditions and geometry before higher order behavior can emerge. Intention, alignment, and governance all operate inside those limits, not outside them.

The question is whether current governance frameworks sufficiently account for these structural constraints, or whether failures in control and stability are better explained by under characterized state spaces.

A concrete treatment of this approach, where toroidal geometry is used as executable boundary conditions in machine learning, is here: https://www.researchgate.net/publication/400051736_TOROIDAL_LIMITS_AS_EXECUTABLE_BOUNDARY_CONDITIONS_IN_MACHINE_LEARNING_V10

What do you think?

More here

Contemporary life is characterized by the rapid rhythm of changes, technological innovations, and AI-driven explorations. Artificial intelligence has transformed the way we use the internet, search information, or get aware of something. It has both strengths and weaknesses, which make us reflect on its efficiency compared to the human mind. From your perspective, can AI-based reasoning surpass human judgment?

Many are currently facing the issue of trust in scientific research conducted using GenAI. Discussions are underway about the need to indicate in the work that generative models were used, a new ethical norm in science is being created, and numerous studies are being conducted on this topic. However, in this case, we unconsciously (or consciously) try to imbue the researcher's tool (GenAI) with the characteristics of a study co-author. But to what extent is this justified and legitimate (I mean the need to indicate the use of AI tools)? What are we trying to achieve? A reduction in the flow of low-quality publications, a decrease in the level of trust in research results (demanding more thorough verification of results), a devaluation of the work of scientists who use AI (do such scientists even exist?)? In my opinion, generative models are simply a research tool, which is currently unreliable (the well-known problem of AI hallucinations). Many researchers (and others) abuse this tool, which ultimately leads to low-quality and even harmful work and incorrect conclusions. In this case, the scientific community's role in verifying research results during the publication approval process is increasing. What does an author's claim that they used GenAI actually indicate? It means they haven't fully understood the conclusions, evidence, reasoning, or code generated by the model. Thus, they are effectively absolving themselves of responsibility for the results of their work ("It wasn't me who made a mistake, it was the model!"; "I didn't fully understand it, but I think it's cool!"). Such works typically have "beautiful packaging," but their content is primitive, empty, or even implausible, hidden behind an abundance of formulas, definitions, and "proofs." In fact, this is the only objective criterion for classifying such works, and it's easily identified by specialists who take the trouble to delve into the work rather than relying on those same models for its analysis. It's important to acknowledge that generative models are a new tool that we need to learn to use, but mentioning them in papers is like mentioning the use of web search or other software 30-40 years ago. What do you think about this?

DJE welcomes original research manuscripts addressing (but not limited to):

Submit your mansucript via https://dibonjournals.com/index.php/dje

AI is being used for preparing papers. Maximum journals has their own rules for that. But it is very difficult to restrain in using AI in writing scientific papers. So how to overcome it? Only declaration of using AI is enough or adding AI as a coauthor will help further?

Dear Language Instructors,

Almost nine months ago, I posted a question about AI and the potential of English language teaching. Now, can we update our discussion about AI and the current practices with specific attention to language skills: (Reading, writing, listening and Speaking)?

Yes, nine months is a short period, but with the advances of AI technology, I am sure that you have experienced many opportunities with AI tools in language skills.

Your discussion is highly appropriated.

Abduljalil

Different industries have seen fast growth of Artificial Intelligence systems which has started discussions about its ethical effects. The expansion of AI systems that affects healthcare and finance as well as social interactions and governance now demands utmost importance for their ethical consistency with human values and societal benefits. The analysis combines modern studies about artificial intelligence ethics to understand its various obstacles along with suggested remedies and future path development within this evolving area. Research presents a complicated situation which includes multiple elements such as raising awareness and educating the public as well as developing ethical frameworks and providing security alongside technical strategies for system implementation together with the cooperation of human and technological values.

A fundamental challenge in AI ethics lies in fostering widespread awareness and understanding of ethical principles among those involved in AI development and deployment. Several studies highlight a significant gap in practitioners' knowledge and awareness of AI ethics [1, 9]. This deficiency underscores the need for proactive initiatives to educate and equip individuals with the necessary tools to navigate the ethical complexities of AI.

To address this gap, researchers have developed and evaluated tools designed to enhance practitioners' understanding of AI ethics [1]. For example, the AI Ethics Quiz was developed to raise awareness and improve knowledge of AI ethics among software practitioners [1]. The results of workshops using the quiz indicated that the quiz significantly improved practitioners' awareness and understanding of AI ethics, and that practitioners found the quiz engaging and reported it created a meaningful learning experience regarding AI ethics [1]. The authors of [1] recommend that software companies and leaders adopt similar initiatives to improve practitioners' understanding of AI ethics. Another study, conducted by [9], found that the majority of AI practitioners had a reasonable familiarity with the concept of AI ethics, primarily due to workplace rules and policies. However, formal education and training was considered somewhat helpful in preparing practitioners to incorporate AI ethics [9].

These findings underscore the importance of integrating AI ethics education into both formal curricula and professional development programs. Initiatives like the AI Ethics Quiz represent promising avenues for improving awareness. However, it is important to note that the effectiveness of such tools may depend on factors such as the target audience, the content of the quiz, and the context in which it is administered [1].

Defining and operationalizing ethical frameworks is a core concern in the field. The literature reflects a diversity of approaches to this challenge, spanning from overarching ethical principles to specific guidelines for the design and deployment of AI systems.

One of the key challenges is to determine which ethical framework to use [2]. The literature presents different branches of applied ethics, such as big data ethics, machine ethics, information ethics, AI ethics, and computer ethics [2]. The authors of [2] provide a clear and brief introduction into normative and applied ethics, describing how they are related to each other and, finally, provide an ordering of the different branches of applied ethics [2].

Several ethical frameworks are proposed to address the ethical challenges in AI, including fairness, transparency, accountability, privacy, and user autonomy [6, 10]. These frameworks often incorporate ethical considerations as core components [6]. For example, the EU has developed Trustworthy Ethics guidelines for Trustworthy AI [11]. However, the literature indicates that there is a challenge of assuring AI ethics with current AI ethics frameworks in real-world applications [3].

The proliferation of ethical principles in AI is not enough to guarantee ethical AI. The development of assurance and verification methods is critical to ensure that AI systems adhere to ethical standards. The concept of AI ethics assurance cases is introduced in the AI ethics assurance [3]. The user requirements-oriented AI ethics assurance case is set up based on three pillars: user requirements, evidence, and validation, and hazard analysis methods used in the safety assurance of safety-critical systems [3]. The authors of [3] also propose a platform named Ethical-Lens (E-LENS) to implement the user requirements-oriented AI ethics assurance approach [3].

Formal methods, such as deontic temporal logic, can be used for specifying and verifying the ethical behavior of AI systems [15]. The authors of [15] propose a formalization based on deontic logic to define and evaluate the ethical behavior of AI systems, focusing on system-level specifications [15]. The effectiveness of this formalization was evaluated by assessing the ethics of the real-world COMPAS and loan prediction AI systems [15]. The formal verification revealed ethical issues in real-world AI applications [15].

Integrating ethical considerations directly into the design and development of AI systems is another important area. This involves exploring how technical mechanisms can be used to promote ethical behavior and mitigate potential harms.

One approach is to leverage game theory to translate human values into ethical AI [12]. The authors of [12] develop a mathematical representation to bridge the gap between the concepts in moral philosophy and AI ethics industry technology standard [12]. As an application, they demonstrate how human value can be obtained from the trust game experiment so as to build an ethical AI [12].

Another approach is to draw on philosophical traditions to inform the development of ethical AI systems [13]. The authors of [13] advocate the notion that Stoic philosophy and ethics can inform the development of ethical A.I. systems [13]. They relate ethical A.I. to several core Stoic notions, including the dichotomy of control, the four cardinal virtues, the ideal Sage, Stoic practices, and Stoic perspectives on emotion or affect [13].

However, it is important to note that technical solutions are not always sufficient to address ethical challenges [14]. The authors of [14] show that the empirical and liberal turn of the production of explanations tends to select AI explanations with a low denunciatory power [14]. Under certain conditions, interpretability tools are therefore not means but, paradoxically, obstacles to the production of ethical AI since they can give the illusion of being sensitive to ethical incidents [14].

The ethical implications of AI are not solely technical; they also involve the complex interplay between humans and machines. This includes examining how AI impacts human autonomy, decision-making, and social interactions.

The literature explores the ethical implications of AI-driven detection technologies in smart homes [10]. This study provides guidelines for ethical design across areas like privacy, fairness, transparency, accountability, and user autonomy [10].

Ethical leadership plays a central role in guiding organizations in facing those challenges and maximizing on those opportunities [6]. A proposed framework for ethical leadership is presented [6].

While formalization and standardization are important, the literature also emphasizes the need for contextual considerations and the limitations of relying solely on formal approaches to AI ethics [7]. The authors of [7] argue that a human ethics of AI based on a pragmatic practice of contextual ethics remains necessary and irreducible to any formalization or automated treatment of the ethical questions that arise for humans [7].

The authors of [8] interrogate the validity of the ETHICS benchmark [8]. Their findings suggest that having a clear understanding of ethics and how it relates to empirical phenomena is key to the validity of ethics evaluations for AI [8].

The field of AI ethics is still nascent but rapidly evolving. A bibliometric analysis of AI ethics literature reveals three phases of development: an incubation phase, making AI human-like machines phase, and making AI human-centric machines phase [4]. The authors of [4] conjecture that the next phase of AI ethics is likely to focus on making AI more machine-like as AI matches or surpasses humans intellectually [4].

The literature highlights several challenges and opportunities for ethical leadership in the age of AI [6]. AI can cause several ethical challenges, including bias in AI algorithms [6]. There are also opportunities for ethical leadership in the age of AI [6]. One of the challenges that AI practitioners face is the incorporation of ethics into AI development [9].

Several key directions emerge from this review. First, there is a need for continued efforts to enhance AI ethics awareness and education. This includes developing and implementing effective training programs for AI practitioners, as well as educating the public about the ethical implications of AI.

Second, further research is needed to refine and operationalize ethical frameworks for AI. This includes developing more specific guidelines for the design and deployment of AI systems in different domains, as well as exploring the use of formal methods to verify and validate ethical behavior.

Third, there is a need to develop technical mechanisms to embed ethical considerations directly into AI systems. This includes exploring the use of game theory, philosophical traditions, and other approaches to promote ethical behavior and mitigate potential harms.

Fourth, more research is needed to understand the complex interplay between humans and machines. This includes examining how AI impacts human autonomy, decision-making, and social interactions.

Fifth, more research is also needed to consider the context and limitations of formalizing ethics.

Finally, there is a need for interdisciplinary collaboration, involving ethicists, computer scientists, policymakers, and other stakeholders [6].

==================================================

AI is quietly shaping both manuscripts and reviewer reports. We know it and want to ask: Does the 'blind' peer review still have any real meaning?

It seems to us, that journals are now taking unacceptably long to review papers and this is slowing down research considerably, with huge impacts for science down the line. We suspect that courtesy of AI, there are more submissions than ever, whilst the number of reviewers remains the same at best (declining at worst). Paying reviewers would obviously help a lot, each and every article discussing peer review says so, but publishers seem to be reluctant to go for this solution, and it’s not that they couldn’t afford to pay, what with their huge APCs, on the one hand, and their very low production costs, on the other. Carrying on as we are risks system shutdown.

The Harbingers Team

In an era where AI tools like large language models provide algorithmic guidance for curriculum design, student assessment, and personalized learning pathways, how might experienced educators integrate these technologies to amplify their leadership roles, such as fostering distributed leadership models or inspiring transformational change in diverse, resource-constrained settings? Drawing from three decades of psychology teaching, I've observed AI's potential to democratize access to ethical direction-setting and value-driven influence, yet concerns arise about algorithmic biases eroding human-centric motivational aspects of leadership.

The EU’s AI Act, U.S. executive orders, China’s algorithm registry—they reflect very different visions of technology and society. Can we develop interoperable standards that respect cultural sovereignty while upholding universal rights (e.g., non-discrimination, privacy)? Or are we heading toward “AI blocs” with incompatible norms?

As models grow more capable—reasoning, planning, even simulating empathy—the risk of misalignment grows. Whose values get encoded? How do we handle pluralism? And can alignment scale beyond narrow tasks to long-term, open-ended goals? This isn’t just a technical problem—it’s deeply philosophical, cultural, and political.

Bettina Gautel and I have tried to address the challenges posed to university teaching by AI systematically and not just reactively.

In our first article, we want to initiate a discussion. You can find the article under the following link:

We have also gone beyond this and tried to research the phenomenon qualitatively, both from the students' and lecturers' perspective. A first publication on this will take some time, as is the case with science.

However, we would be very interested in your experiences in this field, which will keep us all very busy.

I am leading a research project at IIT Madras focused on building a canonical ingredient infrastructure for the Indian market. We are addressing the "linguistic anarchy" found on local food labels by mapping fragmented ingredient variants into a clean, functional taxonomy.

Commitment to Open Data:To ensure this serves as a neutral, public infrastructure good, the final dataset and framework will be released under a CC BY 4.0 license. Our goal is to improve transparency and data interoperability across the Indian food ecosystem.

Responsible AI & Ethics: We are committed to the Responsible Use of AI in this project. We are specifically looking for experts to help us validate:

We have developed a preliminary framework and are seeking high-level validation from Subject Matter Experts to ensure the logic is robust:

Involvement: Due to the pre-publication status of this research, I am not sharing the full dataset publicly yet. I am looking for a one-time, brief review of our Executive Summary and Schema via private message.

All contributing experts will be formally acknowledged in the final publication and the metadata of the public release.

If you have expertise in these areas or can recommend a colleague, please comment below or connect with me directly.

I'd really appreciate anyone with an interest in this area having a look at the platform we have just finished, designed to facilitate assessed conversations, and learning through conversation with AI agents: https://www.hied.ai/

Thanks in advance

I am currently conducting an MRes-level study on Emotional AI and persuasion in political journalism, with a focus on ethical challenges and implications for media trust. I am seeking scholarly literature, theoretical frameworks, or empirical studies that examine:

I would be grateful for recommendations of key journal articles, books, or recent studies (especially from 2018 onwards) that could inform the theoretical or empirical dimensions of this research.

Thank you in advance for any references or insights you may be willing to share.

AI text generator systems are currently taking over plenty of the writing we see around. I would love to find out and hear about tools that are been used to spot it in writing. The concern arises from my lessons: in the field of Language Teaching and Learning passing on the effects to the "machine" will certainly not help students skills development. Maybe it works different for the spoken production but the writing is not there yet! So having a good tool to check the writing I know it's going to help me and consequently my students..

​AI & Society (For General Interest Journals)

​Jobs: How will Artificial Intelligence redefine the concept of "work"? Will human jobs disappear, or will they simply evolve?

​Education: Will AI replace the teacher, or will it serve as a "personal assistant" for every student?

​Creativity: Can a machine possess an "imagination"? And who owns the intellectual property rights to the text and images generated by AI?

​Daily Life: What is the one essential skill every human must learn today to live intelligently in the age of AI?

Ethics & Philosophy (For Cultural Journals)

​Bias: How can we ensure that algorithms do not reproduce the racial or gender biases embedded within our historical data?

​Privacy: In a world where AI analyzes our every move, has the era of "individual privacy" come to an end?

​Control: What are the safety protocols (Alignment Problem) that ensure AI remains a servant to humanity rather than a threat to it?

​Definition: If AI becomes capable of accurately simulating human emotions, does it deserve to be granted "legal rights"?

Recent discussions at Davos, including remarks by Satya Nadella, suggest that AI risks being labeled a “bubble” if its benefits remain concentrated within technology companies rather than translating into broad, measurable productivity gains across industries.

From a research and applied perspective particularly in healthcare what frameworks, metrics, or empirical evidence should we use to assess whether AI is delivering tangible value beyond hype and investment cycles?

Specifically:

I’m interested in perspectives that bridge theory, applied research, and industry deployment, especially where AI adoption intersects with trust, explainability, and regulation.

Hello Dear Everyone,

This issue concerns personalized deregulation. Arguments in favor (Yes) claim that AI can act as an “advocate for innovation” by scanning regulations and proposing targeted, temporary derogations for high-potential projects, which is the most friendly and effective support for SMEs. This accelerates innovation and reduces the oppressiveness of the law. The arguments against (No) warn that the law must be universal and equal for all (the principle of equality); establishing algorithmic exceptions to the rules creates a discretionary and opaque system that can lead to abuse, corruption, and make it difficult for competitors to verify. They believe that AI should optimize the law for everyone, rather than creating ad hoc privileges.

What is your opinion on this topic?

Please reply

I invite you to discuss

Thank you very much. All the best!

I invite you to collaborate on research.

Best regards,

Dariusz Prokopowicz

Dear Colleagues and Friends from RG,

There is a concern that these models, trained on historical data, may unknowingly discriminate against or overburden individuals who are pursuing pioneering social projects, operating in crisis situations, or managing complex structures that inherently exhibit non-standard operational and financial patterns. The implementation of artificial intelligence systems to oversee the public sector is a bold step toward radical transparency, but it takes the discussion of algorithm ethics to a whole new level of state accountability to its citizens. Replacing partially subjective human judgment with precise analysis of large data sets can realistically reduce the scope for corruption, political pressure, or nepotism, which in theory leads to a more equitable distribution of control resources. Nevertheless, the implementation of such systems in public institutions must take into account the fact that social missions often require non-standard actions, which in the eyes of a rigid algorithm may be mistakenly seen as financial anomalies. There is a real risk that fear of a “mathematical verdict” will lead to decision-making paralysis among management, who, instead of looking for innovative ways to solve social problems, will choose safe but less effective courses of action. Such a mechanism could inadvertently punish the most committed individuals who take risks to improve the lives of citizens in emergency situations, such as natural disasters or economic crises. The key challenge, therefore, is to design AI models that can distinguish between creative management and pioneering projects and actual mismanagement or deliberate action to the detriment of the public interest. It is essential to ensure full transparency of the criteria used to avoid a situation where the “black box” of the algorithm becomes a tool of technocratic pressure devoid of human understanding of the context. This process also requires constant human oversight (known as “human-in-the-loop”), which will act as a final safeguard, verifying the machine's indications based on democratic values. Without the appropriate context, these systems may overlook the subtle differences between waste and necessary, long-term investment in social capital. Furthermore, the historical data used to train the systems may contain systemic biases that AI will only perpetuate rather than eliminate, requiring constant auditing of the algorithms themselves. Ultimately, the success of automation of control will depend on finding a happy medium between technological efficiency and the flexibility necessary for the smooth functioning of a modern state. Only by maintaining high ethical standards and social participation in the creation of these tools can AI become a true guardian of integrity, rather than a source of new digital injustice. It is also worth considering the impact of such controls on the general public's trust in the state, which should be built on transparent, not hidden, rules. Reflection on this dilemma should become the foundation of a modern administration that is not afraid of technology but is able to set clear boundaries for its use in public service.

What is your opinion on this topic?

Please reply

I invite you to join the discussion

Thank you very much. All the best!

I invite you to collaborate on research.

Best regards

Dariusz Prokopowicz

Dear Colleagues and Friends from RG,

This issue raises the dilemma between preventive control and stimulating innovation by reducing regulatory stress. Proponents argue that such a solution would radically reduce state oppression and eliminate corruption at the official level, giving entrepreneurs room for business risk. On the other hand, researchers point to the danger of a “black box”—if the algorithm wrongly revokes trust status, the company may have no real recourse against the mathematical decision, raising questions about transparency and the right to human judgment in administrative matters. Considering this issue requires finding a balance between the effectiveness of anomaly detection and the protection of constitutional economic freedoms.

What do you think about this?

Please reply

I invite you to discuss

Thank you very much. All the best!

I invite you to collaborate on research.

Best regards,

Dariusz Prokopowicz