Cosmin GAVRILĂ – ETHICAL AND LEGAL CHALLENGES OF ARTIFICIAL INTELLIGENCE IN INTELLIGENCE: BETWEEN OPERATIONAL EFFICIENCY AND RESPECT FOR FUNDAMENTAL RIGHTS

ABSTRACT:

THE PAPER PRESENTS SOME GENERAL ASPECTS REGARDING THE FIELDS OF USE OF ARTIFICIAL INTELLIGENCE IN THE SPECIFIC ACTIVITIES OF INTELLIGENCE AGENCIES, IN THEIR EFFORTS TO ENSURE THE SECURITY OF STATES.

THE AUTHOR PARTICULARLY HIGHLIGHTS THE OPPORTUNITIES OFFERED BY ARTIFICIAL INTELLIGENCE IN THE INTELLIGENCE PROCESS, ESPECIALLY IN THE STAGES OF INFORMATION COLLECTION AND ANALYSIS, NECESSARY FOR THE CREATION OF THE INTELLIGENCE PRODUCT, WHICH IS THE SUBJECT OF INFORMING THE LEGAL BENEFICIARY, WITH THE AIM OF ENSURING AN ADDITIONAL LEVEL OF KNOWLEDGE FOR THE ADOPTION OF THE STRATEGIC DECISION WITHIN THE EXERCISE OF COMPETENCIES.

COLLATERAL TO THE OPPORTUNITIES OFFERED BY ARTIFICIAL INTELLIGENCE IN THE INTELLIGENCE PROCESS, THE AUTHOR ALSO HIGHLIGHTS SOME ETHICAL ASPECTS CONSEQUENT TO THE USE OF THESE OPPORTUNITIES BY INTELLIGENCE INSTITUTIONS, CONSIDERING THE NECESSITY OF RESPECTING THE FUNDAMENTAL RIGHTS OF CITIZENS, AS THEY ARE STIPULATED IN NATIONAL AND INTERNATIONAL LEGISLATIONS.

KEY WORDS: ARTIFICIAL INTELLIGENCE, INTELLIGENCE, PROCESS, PRODUCT, INFORMATION GATHERING, INFORMATION ANALYSIS, SECURITY

1. GENERAL CONSIDERATIONS

In the era of accelerated digitalization, Artificial Intelligence is redefining the way intelligence agencies operate, providing advanced tools for analysis, surveillance, and national security. The exponential increase in data volume, the complexity of emerging threats, and the need for rapid responses have driven intelligence services to adopt technologies based on machine learning, predictive analytics, and process automation. The use of Artificial Intelligence not only optimizes the collection and processing of information but also allows for the early identification of risks, enhancing the ability of states to respond effectively to security challenges. However, the integration of these technologies raises essential questions regarding transparency, ethics, and data protection, shaping a field that is in continuous evolution and regulation.

As a consequence, in the context of global concerns regarding the assurance of security environment stability, intelligence plays an essential role, both from the perspective of its relation to the intelligence process and from the perspective of the intelligence product [1].

From the perspective of the process, intelligence represents “the set of operations for collecting, filtering, analyzing data and information, and disseminating intelligence products with actionable value to meet the needs of a specific consumer” [2]. Therefore, intelligence is a process of knowledge based on information obtained by specialized structures, which are made available to specific consumers, also called legal beneficiaries, who are established by the specific legislation of each state, in order to provide them with additional knowledge on the issues they manage [3] and to eliminate the uncertainty that characterizes today’s society, marked by a multitude of new security threats that pose unprecedented challenges to intelligence institutions [4].

The intelligence process is characterized by the information cycle, defined by distinct stages such as: formulating the information request by the requester; planning the information collection at the intelligence agency level; collecting primary information; analyzing primary information and creating the intelligence product; disseminating the products to the specific beneficiary [5].

On the other hand, intelligence can be viewed as an “analytical product” resulting from the “intermediate or final activity of analysis” through the processing of raw information data, presented in the form of: periodically prepared information bulletins; informative syntheses; notes; other types of information documents [6].

In both perspectives of approaching intelligence, Artificial Intelligence plays a significant role, which is why intelligence agencies around the world are increasingly using it to enhance their operational capabilities subsumed under the role of ensuring state security. Artificial Intelligence can significantly improve the collection, analysis, and precise dissemination of intelligence activity products [7].

2. THE USE OF ARTIFICIAL INTELLIGENCE IN INTELLIGENCE ACTIVITIES

Referring to the way the intelligence process is defined, we believe that Artificial Intelligence can be primarily used in the stages of collecting and analyzing primary information where it can make significant contributions to intelligence activities and to a lesser extent in planning information collection and disseminating products to the specific beneficiary. Artificial Intelligence plays an essential role in improving the processes of information collection and analysis in the field of intelligence [8]. By automating data collection from open sources and using machine learning algorithms, Artificial Intelligence enables the rapid identification of relevant patterns and trends, thereby reducing the information overload for analysts [9]. Additionally, Artificial Intelligence contributes to the detection of anomalies and emerging threats, facilitating real-time decision-making. However, the integration of Artificial Intelligence into planning and information dissemination processes requires a cautious approach, considering the challenges related to cybersecurity and the ethics of automated decisions [10].

2.1 GUIDELINES FOR USING ARTIFICIAL INTELLIGENCE IN INFORMATION GATHERING

The opportunities offered to intelligence institutions by Artificial Intelligence, in their efforts to achieve knowledge, prevention, and counteraction of security threats, have prompted state entities and international security institutions to adopt coherent legislative measures aimed at regulating the use of these modern technological tools. In this regard, we mention the efforts of the U.S. authorities to use AI in the interest of national security (see Executive Order 13859/2019, which lays the foundation for the American AI Initiative program), or those of the European Parliament, which, on June 14, 2023, adopted the first legislation regulating the use of AI technology at the level of the European Union [7].

In the stage of information gathering by the specialized structures of intelligence agencies, Artificial Intelligence represents a remarkable knowledge resource in exploiting open source information (OSINT), making significant contributions in all four distinct categories of information that can be obtained and used in the intelligence process, namely: Open Source Data (OSD) – raw data from open sources obtained from a primary source; Open Source Information (OSINF) – information from open sources publicly available for mass informing, filtered and primarily validated; Open Source Intelligence (OSINT) – intelligence from open sources intended for a segment of informed beneficiaries, such as military or political decision-makers, to support decision-making; OSINT Validated (OSINT-V) – product of validated and confirmed intelligence from classified sources [11, p. 71-72].

Additionally, Artificial Intelligence represents a source of knowledge for intelligence agencies regarding Social Media Intelligence (SOCMINT) [12], specifically the investigation of social media platforms such as “Facebook,” “YouTube,” blogs, and online discussion channels, obtaining a diverse range of primary information about activities that may be of interest to the mentioned agencies, especially concerning issues related to organized transnational crime such as fraud, identity theft, human trafficking, the proliferation of weapons and weapons of mass destruction, terrorist attacks, cyberattacks on government websites, etc. It also contributes to identifying topics of interest, the motivation behind the actions of the involved actors, their current and future concerns, which are likely to anticipate their actions that could affect certain security interests that are within the purview of intelligence agencies [11, p. 84-87].

The use of dynamic-evolving neural network-type algorithms for exploiting social media sources offers the advantage of providing early warnings regarding possible manifestations of security threat vectors, which can affect the state of equilibrium and security at the national level [13]. Algorithms allow for the rapid extraction and analysis of large amounts of data generated by users on social media platforms such as Facebook, YouTube, blogs, and online discussion channels, thereby identifying trends and abnormal behaviors that could indicate potential threats. Dynamic-evolving neural networks can automatically detect misinformation campaigns or cyber espionage activities by analyzing the sentiments expressed in online posts and identifying irregular patterns in data streams.

For an overview of how the aforementioned algorithms can be used, I will list some technical methods for utilizing Artificial Intelligence in Social Media Intelligence (SOCMINT): web scraping and official APIs for social media platforms such as Facebook, YouTube, Twitter, or Reddit to extract relevant textual content, images, videos, and metadata; the selection and organization of the collected data to eliminate noise and structure the information in a format suitable for analysis; natural language processing (NLP) by applying NLP algorithms to analyze textual content in order to determine the tone and sentiments in texts (positive, negative, neutral), identify themes of interest, motivations for actions and concerns of the involved actors, and identify potential threats or suspicious behaviors; detection of abnormal behaviors through the use of machine learning algorithms to identify unusual patterns that may indicate illegal activities, such as fraud, identity theft, or human trafficking; entity extraction through entity recognition techniques allows for the automatic identification of names of people, organizations, and locations mentioned in online content; detection of topics and trends, models such as Latent Dirichlet Allocation (LDA) are used to identify the main topics discussed on social media; image recognition, convolutional neural networks (CNN) are used to analyze images and detect elements such as weapons, extremist symbols, or illegal activities; audio processing and voice recognition, automatic speech recognition (ASR) techniques transcribe and analyze audio content from videos or live broadcasts; real-time monitoring, the implementation of systems that generate instant alerts upon detecting certain keywords or suspicious behaviors; predictive models, the use of algorithms to anticipate future events based on current trends identified on social media; integrated intelligence platforms, the use of software solutions that centralize and correlate data from multiple sources, providing a unified view of the situation; data visualization, visualization tools such as graphs, maps, and charts are used to present information in an easily interpretable manner that highlights the results of the analysis, thereby facilitating decision-making by intelligence agencies [14, 15, 16].

In the information gathering stage, Artificial Intelligence represents an exceptional knowledge resource, particularly in the exploitation of technical intelligence sources (TECHINT) – Technical INTelligence – which consist of “technical, electronic, optical, physicochemical, mechanical, audio, photo, video means, or results from the combination of these,” used primarily for intercepting, collecting, and processing “sounds, images, or any type of signal or information-bearing medium” [5, p. 60].

Specifically, intelligence agencies use Artificial Intelligence in activities such as: SIGINT Signal Intelligence (SIGINT) – gathering information from sources operating in the electromagnetic spectrum; Communication Intelligence (COMINT) – gathering information by intercepting communications and data transmissions; Electronic Intelligence (ELINT) – the collection of information through electromagnetic transmissions that do not belong to communications; Foreign Instrumentation Signals Intelligence (FISINT) – the collection of information through certain types of electromagnetic signals, using methods different from COMINT and ELINT; Telemetry Intelligence (TELEINT) – the collection of information through the interception, processing, and analysis of foreign telemetry data; Imagery Intelligence (IMINT) – the collection of information from sources specialized in image processing; Photographic Intelligence (PHOTOINT) – collecting information through photo-video surveillance; Measurement and Signatures Intelligence (MASINT) – collection of information through measurements using specific sensors from the analysis of metric, spatial, wavelength, time-dependent, modulation, plasma, and hydromagnetic data; Satellite Intelligence (SATINT) – the collection of technical information using satellites [6, p. 37; 5, p. 94-98].

For Signatures Intelligence (MASINT) activities, Artificial Intelligence makes significant contributions to intelligence agencies in the following areas: Acoustic Intelligence (ACINT) – gathering information by processing acoustic phenomena; Radar Intelligence (RADINT) – gathering information using radar; Nuclear Intelligence (NUCINT) – gathering information from radioactive sources; Infrared Intelligence (INFRAREDINT) – gathering information by exploiting the infrared domain of the electromagnetic spectrum; Laser Intelligence (LASINT) – gathering information by exploiting specific characteristics of laser systems [6, p. 37; 5, p. 99].

Additionally, for Satellite Intelligence (SATINT) activities, Artificial Intelligence makes significant contributions to intelligence agencies in the fields of SIGINT, IMINT, and PHOTOINT – conceptually similar subcategories to those previously presented, with the specification that they operate through satellites – and Geospatial Intelligence (GEOINT) – the collection of information about terrestrial activities and spatial configuration through the exploitation of images and geographic space [6, p. 37; 5, p. 100].

2.2 DIRECTIONS FOR USING ARTIFICIAL INTELLIGENCE IN INFORMATION ANALYSIS

The purpose of intelligence analysis is to interpret a clear and unified vision by following a phased and systematic process of a set of preliminary data or raw information, with the aim of generating qualitative knowledge on an issue that may affect national security [17]. This process involves a rigorous methodological framework, designed to ensure objectivity, relevance, and efficiency in evaluating threats and opportunities in the operational environment [18].

The main purpose of intelligence analysis is the interpretation, correlation, and contextualization of information to enable informed decision-making. The intelligence analysis strategy is based on a structured intelligence cycle, adaptability, the use of emerging technologies (Artificial Intelligence, Big Data), and inter-institutional collaboration to maximize the value of the analyzed information. Intelligence analysis is based on qualitative and quantitative methods, integrated into a structured operational framework, which allows for the generation of actionable insights for decision-makers. Additionally, the methods used include critical and creative thinking, essential for evaluating the relevance of data and observing changes and trends in the manifestation of threats [19].

The final product of intelligence activities is an analytical report. A quality intelligence product must be clear, objective, well-documented, and tailored to the needs of the beneficiary to support them in the decision-making process. Consequently, intelligence analysis plays an essential role in protecting national security, providing decision-makers with fundamental tools for threat prevention. By utilizing robust strategies, advanced methods, and emerging technologies, this process contributes to generating qualitative, actionable, and relevant knowledge for the safety of a state [20].

Specialized literature indicates that the processing of information collected by intelligence institutions is a preliminary stage to information analysis, with the aim of bringing it into accessible forms that intelligence analysts can use to create intelligence products. In this context, Artificial Intelligence is used for recognizing the voices of vectors of interest, detecting dialects or regional linguistic particularities, or identifying certain individuals whose activities represent vectors of security threats. Additionally, Artificial Intelligence is used to establish meta-data regarding the information managed by intelligence agencies, specifically to encode the information based on criteria specific to the agencies, allowing for its search in databases, as well as in the field of information decryption [11, p. 242-243].

Circumscribed to the information analysis stage, intelligence agencies use Artificial Intelligence to analyze large volumes of data that they access or possess (predictive analysis), with the aim of identifying security threats and their manifestation trends, through specific modeling methodologies (see the STRIDE methodology), or to produce intelligence products [11, p. 299].

Another area of application for Artificial Intelligence is augmented analysis, used to enable institutions to support and improve knowledge management practices [21], as well as to optimize analysis and decision-making processes within intelligence institutions, particularly in terms of assessing the credibility and severity of threats and modeling them [11, p. 296].

Additionally, intelligence agencies utilize Artificial Intelligence in data fusion, specifically the combination and integration of information obtained from multiple sources – HUMINT, OSINT, TECHINT – to shape a complete and coherent overview of the investigated issues in general [22], in order to identify trends and action patterns that can contribute to anticipating security risks or action opportunities in the specific domains under investigation [11, p. 301]. It should be noted that such an approach is more than desirable in the activities of modern intelligence institutions, as it contributes to the creation of complex intelligence products.

Recently, intelligence agencies have been concerned with exploiting the opportunities of Artificial Intelligence through predictive analysis, which is based on the use of mathematical and statistical models to identify patterns and trends in the evolution of threats, starting from the data sets managed at a given time by the mentioned agencies, in order to anticipate the trends in the evolution of future events and to offer intelligence agencies opportunities to manage them [23]. Since intelligence agencies may possess large sets of available data in certain investigated domains, primarily obtained from OSINT, Artificial Intelligence can be used to generate “behavioral digital models,” in which the digital “footprint” of a vector under scrutiny can be “combined and correlated” with the history of concerns in order to generate a behavioral profile that can serve for its optimal monitoring [11, p. 311].

From an operational perspective, the use of Artificial Intelligence in intelligence analysis activities is generally carried out through: cognitive automation – specifically assigning certain tasks to the technological infrastructure (developing language models, using ChatGPT algorithms) and behavioral analysis in the case of processing and extracting relevant data [9].

3. ETHICAL AND LEGAL ASPECTS OF THE USE OF ARTIFICIAL INTELLIGENCE IN INTELLIGENCE

The use of Artificial Intelligence in the field of intelligence raises a series of ethical and legal challenges, having a significant impact on the way information is collected, analyzed, and used for national security. As advanced technologies become increasingly integrated into the analytical and operational processes of intelligence agencies, the need for clear regulations arises to balance operational efficiency with the respect for fundamental human rights [24].

From an ethical perspective, the use of Artificial Intelligence raises questions regarding transparency, accountability, and the impact on civil rights, especially concerning population surveillance and automated decision-making. Additionally, the use of Artificial Intelligence systems in intelligence must prevent algorithmic discrimination and ensure adequate human control over decision-making processes. [25].

On the other hand, from a legal perspective, international and national regulations are trying to establish clear limits on the use of Artificial Intelligence, in accordance with data protection legislation, the right to privacy, and cybersecurity [26].

In general, the use of Artificial Intelligence technology in the European space, from a legal perspective, has been regulated since 2018 by the “European Ethical Charter on the Use of Artificial Intelligence in Judicial Systems and Their Environment,” which states that “the use of such tools and services in judicial systems aims to improve the efficiency and quality of justice,” but it must be carried out responsibly, taking into account the need to respect the fundamental rights of individuals, as provided in subsequent European legislation [27]. A year later, in 2019, a group of experts, under the patronage of the European Commission, indicated that the evolution of new emerging digital technologies generated by Artificial Intelligence raises particular issues for clearly establishing a coherent framework of responsibility in the field [7; 28].

In this context, in the year 2023, the European Parliament adopted the legislative framework regulating the use of Artificial Intelligence within the European Union, ensuring safety, transparency, traceability, and human oversight, in order to protect the population from possible undesirable consequences [29].

The particular ethical issues raised by the use of Artificial Intelligence led to the adoption by the European Commission, in February 2025, of a decision to prohibit the following categories of activities within the European Union: banning the use of cameras equipped with real-time facial recognition technologies for the purpose of identifying individuals; social classification based on personal data unrelated to the assessed risk; assessing an individual’s criminal risk based on biometric data; creating databases collected through facial recognition systems by taking images from the internet; recognizing emotions in the workplace or educational institutions; manipulating individuals’ behavior with the help of Artificial Intelligence; exploiting vulnerabilities related to age or disabilities; deducing political opinions or sexual orientation based on biometric data [30].

In these circumstances, international regulations on the use of Artificial Intelligence are applicable to intelligence agencies as well, as they are concerned with adhering to aspects related to: the ethics and confidentiality of specific activities carried out through Artificial Intelligence; the confidentiality and security of data collected and processed through this resource; the explainability of undertaken activities; accountability and bias; the quality of captured, processed, and analyzed data; the collaboration between the human factor and Artificial Intelligence; adverse attacks [11, p. 313].

The use of Artificial Intelligence by intelligence agencies must also comply with the “Recommendations on the Ethics of Artificial Intelligence” adopted at the General Conference of the United Nations Educational, Scientific and Cultural Organization on November 23, 2021, regarding: “respect, protection, and promotion of human rights, fundamental freedoms, and human dignity; proportionality and the prohibition of harm; safety and security; fairness and non-discrimination; transparency and explainability; the right to privacy and data protection; accountability and responsibility” [7; 31].

4. CONCLUSIONS

The use of Artificial Intelligence by intelligence agencies represents an inevitable and necessary evolution in the context of the increasing complexity of threats to national and international security. Artificial Intelligence-based technologies offer significant advantages in the process of data collection, analysis, and interpretation, allowing for a faster and more efficient response to emerging risks.

One of the main benefits of Artificial Intelligence in the field of intelligence is its ability to manage massive volumes of information in real-time, identifying patterns and correlations that might go unnoticed by traditional methods. Additionally, Artificial Intelligence can enhance surveillance and threat anticipation capabilities through predictive analysis and automated processing of data from diverse sources, ranging from social media to cyber communications.
However, the integration of Artificial Intelligence into intelligence activities raises major ethical and legal challenges. Questions regarding the transparency of algorithmic decisions, the protection of civil rights, and the risk of the abusive use of these technologies must be addressed through a clear regulatory framework and effective control and oversight mechanisms. It is essential that intelligence agencies maintain a balance between security and the respect for fundamental rights, ensuring a responsible use of Artificial Intelligence.
In conclusion, Artificial Intelligence represents an essential tool for modernizing and streamlining intelligence activities, but the success of its use depends on the adoption of policies and practices that minimize associated risks. As technology advances, collaboration between governments, security experts, international organizations, and the private sector will be crucial to ensuring the ethical and efficient use of Artificial Intelligence in support of national and global security.

REFERENCES

[1]. Peptan, C. (2019). Information and intelligence in security equation. Analele Universitatii „Constantin Brancusi” din Targu Jiu–Seria Litere si Stiinte Sociale, (02), 39-45.

[2]. Tănase, T., Intelligence-ul modern în Intelligence publicație editată de SRI, nr. 16, iulie – septembrie 2009, pp. 26-28. 

[3]. Peptan, C. (2020). Knowledge and security in modern society through intelligence. Analele Universitatii „Constantin Brancusi” din Targu Jiu–Seria Litere si Stiinte Sociale, (01), 83-89.

[4]. Maior, G., Incertitudine, gândire strategică și relații internaționale în secolulXXI, Editura RAO, București, 2009, p. 24.

[5]. Glad, M., Perja, C., Introducere în studii de Intelligence, Suport dcurs, Universitatea Babeș-Bolyai, Cluj-Napoca, 2024, p. 22.

[6]. Nitu, I., Analiza de intelligence, Editura RAO, Bucuresti, 2012, p. 32.

[7]. Peptan, C., Gavrilă, C., Sîrbu, L., & Mărcău, F. C. (2023). Considerations on the impact of artificial intelligence on terorism. Annals of the Constantin Brancusi University of Targu Jiu-Letters & Social Sciences Series, (1).

[8]. A. D. (17 octombrie 2022). Inteligența artificială. Informații în timp real – Revista Intelligence, Disponibil la https://intelligence.sri.ro/inteligenta-artificiala-informatii-timp-real/. Accessed: 08.02.2025.

[9]. Sfetcu, N. (2023). Integrarea inteligenței artificiale în serviciile de informații – Intelligence Info. Disponibil la https://www.intelligenceinfo.org/integrarea-inteligentei-artificiale-in-serviciile-de-informatii/. Accessed: 08.02.2025.

[10]. Excelență și încredere în inteligența artificială – Comisia Europeană, https://commission.europa.eu/strategy-and-policy/priorities-2019-2024/europe-fit-digital-age/excellence-and-trust-artificial-intelligence_ro. Accesat: 08.02.2025.

[11]. Tănase, T. ș. a., Inteligența Artificială în Intelligence, Editura Concordia, Arad, 2024.

[12]. Omand, D., Bartlett, J., & Miller, C. (2019). Introducing social media intelligence (SOCMINT), în Secret Intelligence (pp. 77-94). Routledge.

[13]. Gagniuc, M., B., Tehnologia ca avantaj competitiv în intelligence. Exploatarea OSINT prin algoritmi de inteligență artificială, disponibil la https://intelligence.sri.ro/tehnologia-ca-avantaj-competitiv-intelligence-exploatarea-osint-prin-algoritmi-de-inteligenta-artificiala/ Accessed: 06.02.2025.

[14]. Browne, TO, Abedin, M. & Chowdhury, MJM O revizuire sistematică a cercetării care utilizează inteligența artificială pentru aplicațiile de inteligență open source (OSINT). Int. J. Inf. Securizat. 23 , 2911–2938 (2024). https://doi.org/10.1007/s10207-024-00868-2

[15]. MS Biagio, R. Acquaviva, V. Mazzonello, E. La Mattina și V. Morreale, „A new SOCMINT framework for Threat Intelligence Identification,” 2021 International Conference on Computational Science and Computational Intelligence (CSCI) , Las Vegas, NV, SUA, 2021, pp. 692-697, doi: 10.1109/CSCI54926.2021.001805

[16]. Blei, D., M., Ng, A., Y., Jordan, M., Latent Dirichlet Allocation, Disponibil la https://www.jmlr.org/papers/volume3/blei03a/blei03a.pdf. Accessed: 09.02.2025

[17/11]. Peptan, C. (2021). A plea for intelligence analysis, Annals of the „Constantin Brancuși” of Târgu-Jiu, Letter ans Social Science, (01), 91-104.

[18]. Sfetcu, N., (2022), Principii ale analizei intelligence în cadrul serviciilor secrete, afaceri și (geo)politică, Intelligence Info. Disponibil la https://www.intelligenceinfo.org/principii-ale-analizei-intelligence-in-cadrul-serviciilor-secrete-afaceri-si-geopolitica/. Accessed: 09.02.2025

[19]. Z. C. (16 august 2022).  Preconcepții si capcane mentale în analiza de informații, https://intelligence.sri.ro/preconceptii-si-capcane-mentale-analiza-de-informatii/. Accessed: 09.02.2025.

[20]. Analiza de intelligence. SRI (Blog). Disponibil la https://www.sri.ro/analiza-de-intelligence. Accessed: 09.02.2025

[21]. Kharisma, D. B., Sudirman, S., & Edi, F. (2023). Current Trend of Artificial Intelligence-Augmented Reality in Science Learning: Systematic Literature Review. Jurnal Penelitian Pendidikan IPA, 9(8), 404-410.

[22]. Meng, T., Jing, X., Yan, Z. și Pedrycz, W. (2020). Un sondaj privind învățarea automată pentru fuziunea datelor. Information Fusion , 57, 115-129.

[23]. Popescu, M. M. (2021). Analiza predictivă a datelor în stratcom în scopul anticipării planurilor strategice. Buletinul Universităţii Naţionale de Apărare» Carol I «, (03), 106-113.

[24]. Sfetcu, N. (2024), Implementarea inteligenței artificiale în securitatea națională, Intelligence Info, https://www.intelligenceinfo.org/implementarea-inteligentei-artificiale-in-securitatea-nationala/.  Accessed: 09.02.2025

[25]. Bidașcă, S. (2023), Etica în inteligența artificială: provocări și perspective, Cunoașterea Științifică, https://www.cunoasterea.ro/etica-in-inteligenta-artificiala-provocari-si-perspective/. Accessed: 09.02.2025

[26]. Rezoluția Parlamentului European din 20 ianuarie 2021 referitoare la inteligența artificială: chestiuni legate de interpretarea și aplicarea dreptului internațional, în măsura în care Uniunea Europeană este afectată în domeniile utilizărilor civile și militare, precum și chestiuni legate de autoritatea statului în afara domeniului de aplicare al justiției penale, https://www.europarl.europa.eu/doceo/document/TA-9-2021-0009_RO.html. Accessed: 09.02.2025

[27]. Carta Etică Europeană în privința utilizării inteligenței artificiale în sistemele judiciare și domeniile adiacente. Disponibil la https://rm.coe.int/ethical-charter-en-for-publication-4-december-2018/16808f699c. Accessed: 07.02.2025.

[28]. Hartmann, J., Jueptner, E., Matalonga, S., Riordan, J., & White, S. (2023). Artificial Intelligence, Autonomous Drones and Legal Uncertainties. European Journal of Risk Regulation, 14(1), 31-48. doi:10.1017/err.2022.15

[29]. Ovidiu, Nahoi. (12 mai 2023). O provocare în fața UE: reglementarea inteligenței artificiale, disponibil la https://www.rfi.ro/politica-social-155762-provocare-fata-ue-reglementare-inteligenta-artificiala. Accessed: 07.02.2025.

[30]. Vidu, V., Comisia Europeană anunţă interzicerea în Uniunea Europeană a opt utilizări ale inteligenţei artificiale, de la supravegherea în masă, la recunoaşterea emoţiilor şi clasificarea socială. Disponibil la https://www.news.ro/externe/comisia-europeana-anunta-interzicerea-in-uniunea-europeana-a-opt-utilizari-ale-inteligentei-artificiale-de-la-supravegherea-in-masa-la-recunoasterea-emotiilor-si-clasificarea-sociala-1922405705002025021021923593. Accessed: 07.02.2025.

[31]. Recomandarea  privind etica inteligenței artificiale. UNESCO.  Disponibil la https://www.cnr-unesco.ro/uploads/media/f1077_recomandari-unesco-ai-site.pdf. Accessed: 07.02.2025.