This study explores the progression of artificial intelligence (AI) systems through the lens of complexity theory, challenging conventional linear projections of advancement toward artificial general intelligence (AGI). We posit the existence of critical points, akin to phase transitions, where increasing system complexity may not lead to greater capability, but rather to performance plateaus or instability. To investigate this hypothesis, we used agent-based modelling (ABM) to simulate the evolution of AI systems, using evaluation benchmark performances as a proxy for complexity. Our simulations modeled the possible characteristics that systems could exhibit when crossing a critical threshold, transitioning from predictable improvement to a regime of erratic, volatile behavior. Practically, we introduced and validated a methodology for detecting these simulated critical transitions algorithmically. We proposed a heuristic Stochastic Gradient Descent-based approach and compared it with established CUmulative SUM (CUSUM) and Lyapunov exponent techniques, to show that different signatures of instability-from abrupt shifts to gradual volatility ramps-can be identified. We contextualized these findings with real-world phenomena, arguing that the empirically observed -"Jagged Capability Frontier" in large language models (LLMs) illustrates the kind of nonlinear performance boundaries that could be sharply accentuated by the onset of criticality. This research contributes not only a novel theoretical framework for understanding potential limits to AI scaling but also a practical, validated methodology for monitoring the systemic stability of AI systems, offering a new dimension to AGI evaluation and safety.

Smart grids rely on advanced metering infrastructure (AMI) for real-time usage data, yet this connectivity introduces vulnerabilities. One critical threat is grid overloading cyberattacks, where an adversary manipulates demand to exceed safe limits and trigger blackouts. Such attacks can harness compromised smart meters to drastically raise the load and destabilize the network. Despite their severity, AMI data manipulation for overloading attacks has received limited attention. We propose a novel anomaly detection framework that combines household and neighborhood-level load predictions. Using deep learning models, our system predicts the electricity usage of each customer, as well as the aggregate community load, allowing detection of sudden deviations that signal an attack. To quantify anomalies, we introduce two metrics: an Abnormality Index (capturing the magnitude of peak deviations) and a Regularity Index (measuring energy consumption pattern consistency). These indices are fed into an ensemble model to detect threats. Crucially, our ensemble model combines multiple classifiers to mitigate adversarial manipulation, making it robust to data poisoning and evasion attempts. Extensive evaluation with a real dataset demonstrates our proposed model achieves a detection rate of 93% and a false alarm rate of 9% in the most challenging baseline conditions, and maintains 81% detection even under the hardest adversarial setting using white-box AutoAttack scenarios, outperforming existing contemporary methods. This integrated solution advances smart grid cybersecurity by combining predictive load modeling, novel anomaly metrics, and robust classification.

The rapid expansion of Internet of Things (IoT) technologies has introduced significant cybersecurity challenges, particularly at the network edge where IoT devices operate. Traditional security policies designed for static environments fall short of addressing the dynamic, heterogeneous, and resource-constrained nature of IoT ecosystems. Existing dynamic security policy models lack versatility and fail to fully integrate comprehensive risk assessments, regulatory compliance, and AI/ML (artificial intelligence/machine learning)-driven adaptability. We develop a novel adaptive edge security framework that dynamically generates and adjusts security policies for IoT edge devices. Our framework integrates a dynamic security policy generator, a conflict detection and resolution in policy generator, a bias-aware risk assessment system , a regulatory compliance analysis system, and an AI-driven adaptability integration system. This approach produces tailored security policies that adapt to changes in the threat landscape, regulatory requirements, and device statuses. Our study identifies critical security challenges in diverse IoT environments and demonstrates the effectiveness of our framework through simulations and real-world scenarios. We found that our framework significantly enhances the adaptability and resilience of IoT security policies. Our results demonstrate the potential of AI/ML integration in creating responsive and robust security measures for IoT ecosystems. The implications of our findings suggest that dynamic and adaptive security frameworks are essential for protecting IoT devices against evolving cyber threats, ensuring compliance with regulatory standards, and maintaining the integrity and availability of IoT services across various applications.

Generative AI models have revolutionized intelligent systems by enabling machines to produce human-like content across diverse domains. However, their outputs often exhibit unpredictability due to complex and opaque internal semantic states, posing challenges for reliability in real-world applications. In this article, we introduce the AI Uncertainty Principle , a novel theoretical framework inspired by quantum mechanics, to model and quantify the inherent unpredictability in generative AI outputs. By drawing parallels with the uncertainty principle and superposition, we formalize the tradeoff between the precision of internal semantic states and output variability. Through comprehensive experiments involving state-of-the-art models and a variety of prompt designs, we analyze how factors such as specificity, complexity, tone, and style influence model behavior. Our results demonstrate that carefully engineered prompts can significantly enhance output predictability and consistency, while excessive complexity or irrelevant information can increase uncertainty. We also show that ensemble techniques, such as Sigma-weighted aggregation across models and prompt variations, effectively improve reliability. Our findings have profound implications for the development of intelligent systems, emphasizing the critical role of prompt engineering and theoretical modeling in creating AI technologies that perceive, reason, and act predictably in the real world.

This study investigated the integration readiness of four predominant cybersecurity Governance, Risk and Compliance (GRC) frameworks - NIST CSF 2.0, COBIT 2019, ISO 27001:2022, and the latest ISO 42001:2023 - for the opportunities, risks, and regulatory compliance when adopting Large Language Models (LLMs), using qualitative content analysis and expert validation. Our analysis, with both LLMs and human experts in the loop, uncovered potential for LLM integration together with inadequacies in LLM risk oversight of those frameworks. Comparative gap analysis has highlighted that the new ISO 42001:2023, specifically designed for Artificial Intelligence (AI) management systems, provided most comprehensive facilitation for LLM opportunities, whereas COBIT 2019 aligned most closely with the European Union AI Act. Nonetheless, our findings suggested that all evaluated frameworks would benefit from enhancements to more effectively and more comprehensively address the multifaceted risks associated with LLMs, indicating a critical and time-sensitive need for their continuous evolution. We propose integrating human-expert-in-the-loop validation processes as crucial for enhancing cybersecurity frameworks to support secure and compliant LLM integration, and discuss implications for the continuous evolution of cybersecurity GRC frameworks to support the secure integration of LLMs.

In response to diverse perspectives on artificial general intelligence (AGI), ranging from potential safety and ethical concerns to more extreme views about the threats it poses to humanity, this research presents a generic method to gauge the reasoning capabilities of artificial intelligence (AI) models as a foundational step in evaluating safety measures. Recognizing that AI reasoning measures cannot be wholly automated, due to factors such as cultural complexity, we conducted an extensive examination of five commercial generative pre-trained transformers (GPTs), focusing on their comprehension and interpretation of culturally intricate contexts. Utilizing our novel “Reasoning and Value Alignment Test,” we assessed the GPT models’ ability to reason in complex situations and grasp local cultural subtleties. Our findings have indicated that, although the models have exhibited high levels of human-like reasoning, significant limitations remained, especially concerning the interpretation of cultural contexts. This article also explored potential applications and use-cases of our Test, underlining its significance in AI training, ethics compliance, sensitivity auditing, and AI-driven cultural consultation. We concluded by emphasizing its broader implications in the AGI domain, highlighting the necessity for interdisciplinary approaches, wider accessibility to various GPT models, and a profound understanding of the interplay between GPT reasoning and cultural sensitivity.

The Internet of Things (IoT) has emerged with Distributed Ledger Technology (DLT) to address existing scalability challenges and improve the trustworthiness of machine-to-machine communication. Among the numerous potential benefits of combining IoT and DLT, Blockchain, a subset of DLT, is a crucial enabler to accelerate secure IoT adoption. Appending a new block to a blockchain, especially in a blockchain-based IoT ecosystem, requires more delay than expected. This delay is one of several issues limiting the broader adoption of blockchain within the IoT domain. To assess this delay, we develop a new comprehensive model to estimate the time required to generate a new block in a blockchain-enabled IoT system. To this end, we develop sub-computation models and compare time consumption associated with the block generation process by conducting an extensive analysis of the following selected IoT layers: device layer, cluster head layer, fog/edge layer, and cloud layer. Our study identifies potential time-consuming steps in adding a new block to a network. Our results demonstrate that the type of blockchain framework and data encryption algorithms could affect the block generation time and that Avalanche, Conflux, Algorand, Polkadot Hyperledger Fabric outperforms Ethereum in terms of block generation time in IoT networks. On the other hand, the blockchain framework does not play a significant role in block generation time for smaller data packets. We also observed the benefit of using 256-bit ECC (elliptic curve cryptography) encryption and the fog layer in IoT networks to enhance the scalability of the block generation process. All in all, our results indicate that the total block generation time varies depending on the selected IoT framework, data encryption algorithm, blockchain type, and key functions of the layers. However, we found that time delays associated with queuing or block size are negligible relative to the other key components of block generation time.

We propose a novel framework, Vacledger, for supply chain traceability and counterfeit detection of COVID-19 vaccines using a blockchain network. It includes four smart contracts on a private-permissioned blockchain network for supply chain traceability and counterfeit detection of COVID-19 vaccine, more specifically to (i) handle the rules and regulations of vaccine importing countries and provide authorization for cross the borders (regulatory compliance and border authorization smart contract), (ii) register new and imported vaccines in the Vacledger system (vaccine registration smart contract), (iii) find the number of stocks that have arrived in the Vacledger system (stock accumulation smart contract), and (iv) identify the exact location of the stock (location tracing update smart contract). Our results show that the proposed system keeps track of all activities, events, transactions, and all other past transactions, permanently stored in an immutable Vacledger connected to decentralized peer-to-peer file systems. We observe no algorithm complexity differences between the proposed Vacledger system and existing supply chain frameworks based on different blockchain types. In addition, based on four use cases, we estimate our model’s overall gasoline cost (transaction or gas price). The Vacledger system empowers distribution companies to manage their supply chain operations effectively and securely using an in-network, permissioned distributed network. This study employs the COVID-19 vaccine supply chain (the healthcare industry) to demonstrate how the proposed Vacledger system operates. Despite this, our proposed approach might be implemented in other supply chain industries, such as the food industry, energy trading, and commodity transactions.

The complete life cycle management of complex equipment is seen as critical to the smart transformation and upgrading of today’s industrial industry. In recent years, digital twin (DT) technology and machine learning (ML) have arisen as emerging technologies. Developing technologies like DT technology and ML in entire battery life cycle management may make each stage of the life cycle more predictable and proactive. We propose a hybrid DT model based on ML that can enhance the performance of an existing DT mathematical model formulated to simulate lithium-ion battery deterioration behavior using DT technology. Firstly, we develop a long short-term memory (LSTM)-based model to forecast the error term of battery capacity enumerated for each charge and discharge cycle from the existing DT model. In this work, we use 18,650 lithium-ion battery discharge data from NASA Ames’ prognostics data repository as our experimental data. The LSTM model is configured with Adam optimizer and the mean absolute error (MAE) loss function. The early stopping criterion is also employed as a regularization technique to overcome model overfitting. Secondly, we develop our proposed hybrid DT by integrating both the existing DT and the LSTM model. Thirdly, we formulate an empirical mathematical model, which allows us to better replicate behavior of battery degradation of any lithium-ion battery. Finally, we evaluate the performance of the proposed hybrid DT in terms of the MAE metric. Compared with the existing model, our proposed model reduces the error of battery capacity during the entire degradation period by 68.42%.

Social media has beneficial and detrimental impacts on social life. The vast distribution of false information on social media has become a worldwide threat. As a result, the Fake News Detection System in Social Networks has risen in popularity and is now considered an emerging research area. A centralized training technique makes it difficult to build a generalized model by adapting numerous data sources. In this study, we develop a decentralized Deep Learning model using Federated Learning (FL) for fake news detection. We utilize an ISOT fake news dataset gathered from "Reuters.com" (N = 44,898) to train the deep learning model. The performance of decentralized and centralized models is then assessed using accuracy, precision, recall, and F1-score measures. In addition, performance was measured by varying the number of FL clients. We identify the high accuracy of our proposed decentralized FL technique (accuracy, 99.6%) utilizing fewer communication rounds than in previous studies, even without employing pre-trained word embedding. The highest effects are obtained when we compare our model to three earlier research. Instead of a centralized method for false news detection, the FL technique may be used more efficiently. The use of Blockchain-like technologies can improve the integrity and validity of news sources.

Sharding is considered to be the most promising solution to overcome and to improve the scalability limitations of blockchain networks. By doing this, the transaction throughput increases, at the same time compromises the security of blockchain networks. In this paper, a probability distribution model is proposed to analyze this trade-off between scalability and security of sharding-based blockchain networks. For this purpose hypergeometric distribution and Chebyshev's Inequality are mainly used. The upper bounds of hypergeometric distributed transaction processing and failure probabilities for shards are mainly evaluated. The model validation is accomplished with Class A (Omniledger, Elastico, Harmony, and Zilliqa), and Class B (RapidChain) sharding protocols. This validation shows that Class B protocols have a better performance compared to Class A protocols. The proposed model observes the transaction processing and failure probabilities are increased when shard size is reduced or the number of shards increased in sharding-based blockchain networks. This trade-off between the scalability and the security decides on the shard size of the blockchain network based on the real-world application and the blockchain platform. This explains the scalability trilemma in blockchain networks claiming that decentralization, scalability, and security cannot be met at primary grounds. In conclusion, this paper presents a comprehensive analysis providing essential directions to develop sharding protocols in the future to enhance the performance and the best-cost benefit of sharing-based blockchains by improving the scalability and the security at the same time.

Despite sentiment analysis or contextual mining of text that recognizes and extracts subjective information from a source, it is considered necessary to estimate human behavior. A hashtag is a metadata tag used to classify data into a category. However, there has been little discussion on segmenting hashtags so far. We propose an algorithm to segment hashtags by optimizing computation time. We create candidates according to a given corpus, containing 1-gram (unigram) and 2-gram (bigram) data. The proposed algorithm allows a reduction in the computation time of generating segments by limiting the candidates in a given corpus. The fewer candidates there are, the shorter the calculation is, leading to a decreased duration. In this study, we gather food-related unstructured tweets (N = 951,255) from Twitter. Our results demonstrate that the proposed algorithm allows a computation time reduction of 29.7%. However, if the segment could not be found with the proposed algorithm, the original method for hashtag segmentation, which includes identifying all possible candidates, is used as a fallback method. The proposed approach improves the hashtag segmentation technique, minimizing computation time, which could be utilized in real-time tweet analysis. The result of our study shows that the trend of sentiments for both raw data and segmented data is similar, which also verifies the method's accuracy. This study's discoveries uncover that, despite the fact that computers are getting faster, computational resources should be utilized effectively. Our work also provides a data collection model for future surveys, which could also shorten the data retrieval process with multi-threading programming concepts.

This letter establishes the similarity between the sensor scheduling problem and the handoff (i.e., base station assignment) problem in cellular networks. A mobile user behavior is then modelled by a Hidden Markov Model (HMM). The handoff problem is formulated as an optimization problem of base station scheduling that minimizes a cost function that involves the HMM state estimation error and base station measurement costs. The optimization problem can be solved using algorithms known as partially observed Markov decision processes.

This paper proposes a new call quality measure based on mobile signal strength measurements to evaluate performance of handoff algorithms in wireless cellular networks. The proposed measure allows the quantification of the impact of the handoff algorithms of performance. Using the proposed measure we compare existing handoff algorithms to identify the trade-off between signal quality and required number of handoffs. Our results indicate that a handoff method based on a threshold with 2 dB hysteresis provides better performance compared to the conventional wisdom of 3 dB hysteresis. We provide a benchmark value for handoff algorithms based on an off-line heuristic method using the new measure. Our benchmark shows that there is substantial room for improvement of the existing handoff algorithm