Artificial Intelligence Governance in Brazilian Organizations: an Analysis in Light of PL 2.338/2023

Aline Fernanda de Oliveira; Julia Pinto Komka

DOI: 10.22167/2675-6528-202602618

Article derived from a Course Conclusion Work (TCC), with content based on the student’s original work and adapted to the editorial format of the E&S Magazine with the support of the ResumeAI tool, an artificial intelligence solution developed by Instituto Pecege for textual synthesis and organization.

Summary

The growing adoption of artificial intelligence (AI) in organizations has intensified the need to understand whether its incorporation has been accompanied by governance, risk management, and regulatory preparedness mechanisms compatible with the technology’s challenges and the requirements of Bill No. 2,338/2023. The degree of perception, preparedness, and engagement of Brazilian organizations regarding AI governance, in light of the national regulatory proposal, was analyzed. The research was characterized as applied and descriptive, with a quantitative and qualitative approach, using a documentary and field survey design. A structured questionnaire was applied to 58 respondents, and the data were subjected to descriptive statistical and content analysis, complemented by the analysis of legal documents and references on AI governance. The results indicated that 79.3% of participants worked in organizations that used AI, with adoption concentrated in intermediate maturity stages. Relevant recognition of the risks associated with AI was observed, especially in privacy, data protection, and human supervision, but the implementation of formal governance mechanisms remained limited. Regarding regulatory preparedness, 48.3% did not know about adaptation plans for Bill No. 2,338/2023, and only 15.5% indicated structured initiatives. The study highlighted the need to strengthen processes, risk management, training, and regulatory monitoring to align AI adoption with more mature governance and compliance structures.

Keywords: compliance; ethics; risk management; human oversight; digital transformation.

1. Introduction

The growing adoption of artificial intelligence (AI) systems in organizations has intensified debates about their ethical, social, and legal impacts. When used strategically, AI can generate competitive advantages by improving processes, optimizing decisions, and creating sustainable value (Fernandes et al., 2025). In parallel, these technologies transform work dynamics and require new professional skills, demanding continuous adaptation and individual protagonism in production processes (Gabriel, 2025). Thus, AI transcends the role of a technological tool, reconfiguring productive and human structures within companies.

AI systems are defined as autonomous technologies that process data and generate results with impact in virtual or real contexts (Brasil, 2024). In practice, AI is based on machine learning techniques, such as deep neural networks, which identify patterns in large volumes of data to generate predictions and decisions, without reproducing human reasoning (Silva et al., 2023).

In recent years, discussions about the ethical limits of AI have intensified due to negative impacts associated with algorithmic biases in systems used by public and private organizations (Silva et al., 2023). These controversies broaden the perception that technological advancement must be accompanied by control mechanisms and safeguards for fundamental rights. Although AI brings gains in efficiency and innovation, it also introduces significant risks, such as algorithmic biases, unclear automated decisions, and potential violations of fundamental rights. These challenges demand solid governance and compliance structures that ensure the ethical and responsible use of these technologies (Brasil, 2024).

Given the risks associated with the use of algorithmic systems, the need for governance structures oriented towards risk management stands out, capable of mitigating impacts, ensuring transparency, and promoting the accountability of the agents involved throughout the lifecycle of automated decisions (Koene et al., 2019). In this sense, Amorim and Silva (2025) understand AI governance also as a coordination strategy between different agents, given the transversal nature of the technology. For Silva et al. (2023), technological governance constitutes an intermediate layer between ethics and law, allowing for more agile responses than legal regulation, through the definition of policies, standards, and procedures guiding the responsible use of information technologies. Governance acts as a bridge between ethical principles and normative application, offering control instruments adaptable to the speed of technological innovation.

In Brazil, Bill nº 2.338/2023 establishes general rules for the responsible governance of AI systems, focusing on human centrality and the protection of fundamental rights. The text proposes principles of transparency, due diligence, accountability, risk prevention, and mitigation (Brazil, 2023). The legislative proposal was developed based on international references, adopting part of the guidelines from the European Union’s Artificial Intelligence Act, especially in the risk-based approach and the provision of governance mechanisms and algorithmic impact assessment (Brazil, 2024). The legislative proposal also introduces the mandatory preliminary and algorithmic impact assessment, reinforcing the need to identify, assess, and mitigate risks arising from the use of AI (Brazil, 2023).

Bill No. 2,338/2023 highlights the importance of human supervision in high-risk AI systems, ensuring that automated decisions remain under human control and that any risks or impacts on rights are identified and mitigated (Brazil, 2023). This principle reinforces the need for technology to remain subordinate to critical capacity and human intervention, consolidating the centrality of the person as a structuring value for future regulation. In the context of this research, corporate engagement is understood as the degree of institutional, operational, cultural, and regulatory commitment to adopting AI governance and compliance practices, encompassing the involvement of senior management and the compliance, risk, and technology areas (Brazil, 2023).

Digital governance encompasses a broader scope than compliance, involving ethical, social, and technological dimensions that influence organizational performance in a context of continuous digital transformation (Silva et al., 2023). Understanding this interdependence between governance, compliance, and innovation is essential to assess how companies prepare to meet future legal and ethical requirements associated with the use of AI. Given this scenario of rapid technological and regulatory evolution, it becomes crucial to analyze the capacity of Brazilian organizations to align the adoption of artificial intelligence with robust governance and compliance frameworks. This work aimed to analyze the degree of perception, preparedness, and engagement of organizations regarding artificial intelligence governance, in light of the regulatory proposal of Bill No. 2,338/2023.

2. Material and Methods

The present research was characterized as applied, seeking to generate knowledge for the solution of practical problems and to improve governance and compliance in AI in the corporate environment (Bigaton et al., 2024). Additionally, it was descriptive in nature, with the purpose of identifying and detailing behaviors, perceptions, and practices observed in organizations. This approach allowed for a portrayal of reality in the face of the regulatory advancement of Bill No. 2,338/2023, without interference in the phenomena studied (Gil, 2008).

A mixed methodological approach was adopted, combining quantitative and qualitative elements, as advocated by Bigaton et al. (2024). The quantitative aspect was used to measure the degree of perception, preparedness, and engagement of companies regarding AI governance, using data from a structured questionnaire. The qualitative dimension, in turn, focused on documentary analysis and the interpretation of legal provisions and theoretical frameworks, deepening the understanding of the phenomenon.

The research design comprised documentary studies and field survey (survey). The documentary research involved the analysis of materials that had not received prior analytical treatment, such as laws, regulations, and institutional documents (Bigaton et al., 2024). The empirical object of the study was the perception, preparedness, and engagement of Brazilian organizations regarding artificial intelligence governance.

The documentary stage involved reading and interpreting Bill No. 2,338/2023 and the ANPD’s Substitute Bill (2024). These documents were complemented by international references on the governance of algorithmic systems, notably Koene et al. (2019), which address transparency, accountability, risk management, and oversight. The objective was to identify the legal principles and mechanisms that guide compliance and human supervision in the use of AI systems.

The field survey was conducted using a structured questionnaire, developed specifically for this research by the author. The instrument was not reproduced or adapted from others, but based on the theoretical and normative framework of the study, including Bill No. 2,338/2023, the ANPD Substitute (2024), and the contributions of Silva et al. (2023), Fernandes et al. (2025), and Gabriel (2025).

The structured questionnaire, composed of 30 questions and the Free and Informed Consent Form (TCLE), covered five thematic blocks. These included respondent profile, AI adoption stage, risk perception and governance, compliance and preparation for Law No. 2,338/2023, and organizational engagement and culture, in addition to an open-ended question. Before final application, the questionnaire underwent a pilot test to verify clarity and comprehension.

The research participants were professionals from diverse backgrounds and sectors, with a preference for those who had interfaces with governance, risk, compliance, innovation, technology, or corporate management. Data collection occurred online, between December 2025 and March 2026. At the end of the period, 58 valid responses were obtained, a number that fit within the initial estimate of 50 to 80 responses, defined by exploratory criteria and feasibility.

To ensure research ethics, the Consent agreement (TCLE) was applied to all participants. Responses were collected anonymously and confidentially, and used exclusively for academic purposes. No information allowed identification of the participant or organization, ensuring voluntariness and the possibility of interrupting participation at any time, without the need for justification.

The quantitative data collected through the questionnaire were submitted to descriptive statistical analysis, according to the guidelines of Bigaton et al. (2024). This technique was employed to identify patterns, frequencies, and possible cross-tabulations between variables, allowing for the interpretation of the degree of companies’ engagement and its relationship with the principles of the regulatory proposal for AI.

In a complementary way, a content analysis of the open-ended responses was carried out, following Bardin’s (2011) methodology. This technique allowed for the exploration of the material and the identification of recurring categories, extracting relevant qualitative meanings and interpretations from the participants’ discourse. The quantitative and qualitative results were integrated and confronted with the study’s theoretical and legal framework.

During the preparation of the work, ChatGPT (OpenAI) was used as a support tool. Its application was limited to formatting, spelling and grammar correction, paraphrasing assistance, and abstract translation. All content was reviewed and adapted by the author, who assumed full responsibility for the final work.

3. Results and Discussion

The analysis of the collected data revealed a complex panorama regarding artificial intelligence governance in Brazilian organizations, indicating that, although the adoption of AI is a consolidated reality, the structuring of governance mechanisms, risk management, and regulatory preparation is still in its initial stages. The findings highlight a disconnect between technological advancement and the formalization of control practices, which could compromise compliance with the future requirements of Bill No. 2,338/2023. The research sought to analyze the degree of perception, preparedness, and engagement of organizations concerning AI governance, in light of the national regulatory proposal, and the results obtained provide important subsidies for this understanding.

Sample profile and initial characterization

The study sample consisted of 58 respondents, with a predominance of the female gender, representing 74.1% of the total, and a significant geographical concentration in the Southeast region, with 70.7% of participants. The South and North regions showed lower representativeness, with 12.1% and 10.3%, respectively. This concentrated regional distribution suggests a limitation in generalizing the results to other areas of the country, although the respondents’ profile was diversified in terms of experience and sector of activity.

Regarding professional experience, a predominance of intermediate and high levels was observed, with 27.6% of respondents in the 11 to 20 years range, 24.1% between 3 and 5 years, and 19.0% between 6 and 10 years. The ranges of lower and higher experience, with up to 2 years and over 20 years, represented 13.8% and 15.5%, respectively. This profile indicates that the sample is mostly composed of professionals with relevant experience in the organizational environment, which strengthens the consistency of perceptions on the research topic.

Sectoral diversity was notable, with the “other” category accounting for 22.4% of the sample, and sectors such as industry, health, non-financial services, and technology and innovation, each with a 12.1% share. Financial services represented 10.3% of respondents. This sectoral heterogeneity is relevant, as different segments may present distinct levels of regulatory maturity and risk exposure, influencing how AI is adopted and managed, as pointed out by Silva et al. (2023).

Regarding the professional profile, technical or specialized positions were predominant, with 58.6% of respondents, followed by leadership positions, which represented 34.5%. This composition indicates an audience with analytical capacity and potential involvement in organizational decisions. The most represented areas of activity were Legal (25.9%) and Compliance/Governance (22.4%), followed by Information Technology/Data (15.5%) and Risk Management/Internal Controls (12.1%).

The significant presence of professionals from the Legal, Compliance, and Governance areas in the sample is consistent with the research objective, as these respondents tend to have greater familiarity with control, compliance, and risk management practices, especially in the context of AI. However, this predominance may also influence the results, increasing the perception of maturity or criticality regarding the topic, which should be considered in the interpretation of the findings.

Adoption of artificial intelligence in organizations

The results indicated that artificial intelligence is already a reality in most of the analyzed organizations, with 79.3% of respondents affirming its use, while 19.0% have not yet adopted it. This data highlights the consolidation of AI as a relevant tool in the corporate environment, reinforcing the trend of increasing adoption of AI-based technologies in digital transformation contexts, as discussed by Gabriel (2025), although its incorporation is not homogeneous among different organizations.

Regarding the stage of AI adoption, a concentration was observed at intermediate maturity levels. It was found that 43.1% of organizations are in a partially implemented stage, and 32.8% are in an experimental or pilot phase. Only 15.5% of respondents indicated broad and integrated use of AI in organizational processes. This scenario suggests that, although the technology is present, its strategic and full incorporation is still evolving, demanding greater formalization and integration.

Regarding internal guidelines for the ethical use of AI, 43.1% of organizations stated they have related policies or guidelines, while 32.8% indicated partial adoption and 19.0% reported non-existence. Additionally, 5.2% of respondents were unsure about the existence of these guidelines. This distribution of responses suggests that the formalization of internal policies is largely aligned with the adoption of governance guidelines, but at different maturity levels, with significant gaps in some organizations.

The human validation of automated decisions was a relevant point. It was found that 43.1% of respondents indicated that it always occurs, and 22.4% stated that it frequently occurs. However, 17.2% of participants did not know how to evaluate this practice, which may indicate a lack of clarity or internal communication about these processes. This “don’t know how to evaluate” percentage suggests possible gaps in transparency or formalization of human oversight mechanisms, one of the pillars of AI governance.

These findings corroborate the literature that points out that the adoption of AI systems tends to occur more rapidly than the implementation of governance and control structures (Silva et al., 2023). Bill No. 2,338/2023 establishes the need for human supervision and governance mechanisms throughout the AI lifecycle (Brazil, 2023), which highlights that the absence of these practices can represent future non-compliance risks. The mismatch between technological advancement and the formalization of governance practices may compromise adherence to emerging regulatory requirements.

Risk perception and governance mechanisms

The analysis of the level of organizational concern regarding the risks associated with AI revealed a priority focus on aspects directly related to data protection. Data privacy and protection were indicated with the maximum level of concern (level 5 on a Likert scale) by 56.9% of respondents, evidencing a high sensitivity to the topic. This finding reflects the growing importance attributed to information security in a scenario of intensive use of digital technologies.

Another factor of relevant concern was human supervision in AI decisions, with 56.9% of respondents indicating high levels of concern (levels 4 and 5 on the Likert scale). In contrast, the risk of algorithmic biases showed a greater concentration in intermediate levels, with 39.7% of responses at level 3. This result suggests that risks more directly associated with regulation and accountability, such as privacy and human supervision, are more recognized by organizations than risks of a technical nature, such as algorithmic biases, which still present lower maturity of understanding.

Regarding governance mechanisms, only 36.2% of respondents affirmed the existence of an area responsible for supervising the use of AI. A significant portion, 29.3%, indicated the non-existence of such an area, and 25.9% pointed to partial adoption. This result highlights that the definition of responsibilities is not yet fully institutionalized within organizations, suggesting that AI governance is not structurally integrated with traditional corporate governance, often remaining dispersed among areas or nonexistent.

Regarding the formal risk assessment, only 25.9% of respondents stated they perform this type of practice, while 27.6% indicated its absence and 24.1% did not know how to assess. This scenario may reflect both the absence of structured processes and the low internal visibility of existing mechanisms, reinforcing a misalignment between the perception of risks and the organizational capacity to structure effective governance and control mechanisms. This gap compromises the effectiveness of compliance and risk management programs applied to AI.

Regarding the existence of mechanisms for correcting or suspending systems in case of failures, 32.8% of respondents indicated the presence of these mechanisms, while 27.6% pointed to partial adoption and 24.1% did not know how to assess. The absence or lack of knowledge of these mechanisms can represent a relevant risk, especially in scenarios of systemic failures or automated decisions with significant impact. These findings are aligned with the literature that points out that the governance of algorithmic systems depends on structured mechanisms for accountability, transparency, and monitoring (Koene et al., 2019).

Compliance and preparation for PL No. 2,338/2023

The results indicated that the regulatory maturity of organizations is still under development, especially regarding the monitoring of initiatives related to AI. It was observed that 39.7% of respondents did not know whether their organizations follow discussions about Bill No. 2,338/2023, while 24.1% stated that they do not carry out this monitoring. This data reveals significant levels of unawareness or absence of structured monitoring, which may compromise the capacity for regulatory anticipation.

Regarding the preparation for the eventual entry into force of the regulation, should Bill No. 2.338/2023 be approved, the data reinforce a scenario of low maturity. It was found that 48.3% of respondents did not know about the existence of adaptation plans, and only 15.5% stated that their organizations have structured initiatives in this regard. This result shows that the topic has not yet been systematically incorporated into the organizational agenda, indicating a predominantly reactive stance by organizations, in contrast to good compliance practices that recommend a preventive and anticipatory approach.

Regarding internal training on AI ethics and governance, a relatively balanced distribution was observed among organizations with no training (36.2%), those that claim to have it (34.5%), and those indicating partial implementation (27.6%). This result suggests that, although the topic is already present in some initiatives, there is not yet a consolidation of structured competency development programs. The absence of structured training may limit the effectiveness of governance policies and mechanisms, as the understanding of principles depends on the development of organizational competencies.

The general perception of the organizational readiness level reinforces this diagnosis. The majority of respondents classified readiness as medium (36.2%) or low (32.8%), while only 20.7% indicated a high level of preparation. This scenario highlights that most organizations do not yet consider themselves fully capable of meeting the demands of future regulation. These findings demonstrate a significant gap between current organizational practices and the principles established by Bill No. 2,338/2023, which provides guidelines such as transparency, due diligence, accountability, risk mitigation, and human oversight throughout the AI lifecycle (Brazil, 2023).

From this perspective, the absence of regulatory monitoring, structured planning, and continuous training can compromise organizations’ ability to adapt to regulatory requirements. The literature indicates that the incorporation of governance practices and preparation for regulatory demands are relevant factors for risk reduction and the sustainability of emerging technology use (Fernandes et al., 2025). The observed low maturity suggests that many organizations still operate reactively, increasing vulnerability to future legal requirements.

Engagement and organizational culture

The analysis of engagement and organizational culture in relation to AI indicated an intermediate level of maturity, reflected in the predominance of medium-level engagement perceptions among employees. This scenario suggests a still gradual incorporation of AI governance practices in the organizational environment. The effectiveness of AI governance depends on the articulation between leadership, training, and institutionalization of practices, as highlighted by Silva et al. (2023).

Regarding the interdisciplinary treatment of the subject, it was observed that AI is approached in an integrated manner across areas such as business, risks, and technology in a relevant part of the organizations, being classified as frequent by 34.5% of respondents and always by 17.2%. On the other hand, an expressive portion indicated that this integration occurs only occasionally or in a limited way, with 22.4% pointing out that it happens sometimes and 24.1% indicating low recurrence (rarely or never). This suggests that the integration between areas is not yet fully consolidated, and AI governance faces challenges of organizational transversality, often being treated in a fragmented manner.

The perception of the role of senior management is predominantly positive, with 55.2% of respondents indicating that there is institutional incentive for the ethical and responsible adoption of AI, and 25.9% pointing to partial incentive. This result demonstrates that the topic is already present on the strategic agenda of a relevant part of the organizations, albeit at different levels of maturity. The incentive from senior management, although relevant, is not sufficient to guarantee the effective internalization of governance practices.

Despite institutional encouragement, employee engagement is not reflected in the same proportion. The majority of respondents classified this engagement as medium (60.3%), while 22.4% indicated a high level of engagement. This difference highlights that the internalization of AI governance practices is not yet fully widespread among operational levels. The difficulty in converting strategy into practical execution presents itself as one of the main challenges for the implementation of governance practices, especially in digital transformation scenarios.

This mismatch between strategic direction and organizational engagement can be interpreted in light of the literature on digital transformation, which points out that the adoption of new technologies requires not only technical investments but also the development of skills, organizational culture, and continuous learning processes (Gabriel, 2025). Bill No. 2,338/2023 establishes guidelines such as transparency, responsibility, and human supervision, whose effectiveness is related to the organizations’ capacity to internalize these principles, which presupposes not only formal structures but also cultural engagement and organizational alignment (Brazil, 2023).

Challenges and qualitative analysis

The analysis of challenges related to the implementation of AI governance practices showed that the main barriers faced by organizations are associated with structural and organizational factors, rather than technological limitations per se. Respondents pointed to the lack of technical knowledge as the biggest challenge, with 39 mentions, followed by the absence of specific standards, with 27 mentions. Budgetary limitations were cited by 19 respondents, and lack of leadership support by 15, in addition to other difficulties in smaller proportions.

This result reinforces that the main challenges are not related to the technology itself, but to the organizational capacity to structure and operationalize governance models. In addition, the analysis of the responses to the open question about measures to strengthen AI governance and compliance revealed coherence between the identified challenges and the proposed solutions. The statements indicated as main actions the realization of training, with 25 mentions, the definition of policies and procedures, with 14 mentions, and the strengthening of risk management, with 10 mentions.

Other initiatives suggested by respondents included promoting transparency, security, and regulatory enhancement. The correspondence between challenges and solutions highlights a maturity in respondents’ perception of organizational needs. The qualitative analysis of open-ended responses, conducted using content analysis techniques (Bardin, 2011), allowed for the identification of recurring categories that deepen these findings, emphasizing the need for greater regulatory clarity, internal training, definition of responsibilities, and strengthening of control and monitoring mechanisms.

Amorim and Silva (2025) highlight that AI governance is not limited to the adoption of normative guidelines, also requiring mechanisms for evaluation, continuous supervision, and institutional adaptation. These results align with the literature that points out that AI governance depends on the integration of technical, organizational, and normative aspects, with the definition of policies, processes, and responsibilities throughout the AI systems’ lifecycle being fundamental (Silva et al., 2023). Bill No. 2.338/2023 emphasizes the need for risk assessment, transparency, accountability, and human oversight, elements directly reflected in both the challenges and solutions pointed out by the respondents.

In general, the identified challenges do not predominantly focus on technological limitations, but rather on the capacity of organizations to structure governance models that ensure their responsible use in compliance with ethical and regulatory principles. This result confirms that AI governance is, to a large extent, an organizational and institutional challenge, and not just a technological one. The barriers to AI governance are directly related to the absence of training, clear guidelines, and organizational structuring, being consistent with the measures suggested by the respondents, who point to the need for skills development, formalization of policies, and strengthening of risk management. Overcoming these challenges depends on an integrated approach, involving leadership, organizational culture, training, and regulatory alignment.

In summary, the research results indicate that, although artificial intelligence is widely present in Brazilian organizations, the maturity of governance and compliance is still intermediate. A relevant recognition of the risks associated with AI was observed, especially in privacy, data protection, and human oversight, but the implementation of formal governance mechanisms and regulatory preparation for Bill No. 2,338/2023 remain limited. The study highlighted the need to strengthen processes, risk management, training, and regulatory monitoring to align AI adoption with more mature governance and compliance frameworks, responding to the objective of analyzing the perception, preparedness, and engagement of Brazilian organizations.

4. Conclusion

This study sought to analyze the degree of perception, preparedness, and engagement of Brazilian organizations regarding artificial intelligence governance, in light of the regulatory proposal of Bill No. 2,338/2023. It was found that artificial intelligence is already widely used in organizations, with most respondents indicating its adoption. However, it was observed that this incorporation is concentrated in intermediate maturity stages, without full integration into organizational processes. A significant recognition of the risks associated with AI was identified, especially concerning privacy, data protection, and the need for human supervision in automated decisions. Nevertheless, the implementation of formal governance mechanisms, such as dedicated oversight areas and structured risk assessments, remained limited. Regulatory preparedness for Bill No. 2,338/2023 proved to be incipient, with a considerable portion of respondents unaware of adaptation plans or indicating initiatives that are not yet structured. This scenario highlights a disconnect between the advancement of technological adoption and the formalization of governance and compliance practices.

The main contribution of this work lies in presenting empirical evidence on the maturity of AI governance in Brazilian organizations, reinforcing the relevance of the topic for compliance, ESG, and regulatory preparation agendas. The findings indicate that artificial intelligence governance constitutes a predominantly organizational and institutional challenge, depending on the structuring of processes, the development of competencies, and alignment with future regulatory requirements. However, the study has limitations, such as the regional concentration of the sample in the Southeast region and the predominance of professionals from the Compliance, Legal, and Governance areas, which may influence the perception of risks and control practices. Furthermore, the non-probabilistic nature of the sample restricts the generalization of the results. For future studies, it is suggested to investigate the effectiveness of training programs and the implementation of formal governance policies, as well as to explore the perception in other regions of the country, in order to deepen the understanding of organizations’ adaptation to AI regulation.

Bibliographic References

AMORIM, A. B.; SILVA, R. C. 2025. Governança de IA no Brasil: proposta de modelo para equilíbrio tecnológico e econômico. Trabalho de Conclusão de Curso em Sistemas de Informação. Universidade Presbiteriana Mackenzie, São Paulo, SP, Brasil. Disponível em: https://dspace.mackenzie.br/items/d189236a-a06b-4840-9302-b023597ebe7b. Acesso em: 19 ago. 2026.

ANPD, 2024 [Referência completa não encontrada no documento original]

BARDIN, L. 2011. Análise de conteúdo. Edições 70, Lisboa, Portugal

BIGATON, A.; GOBBI, M.; OLIVEIRA, D.; GOMES, J. 2024. Metodologias de pesquisa. PECEGE/USP-Esalq, Piracicaba, SP, Brasil.

BRASIL. 2023. Projeto de Lei nº 2.338, de 2023: dispõe sobre o uso da Inteligência Artificial no Brasil e dá outras providências. Câmara dos Deputados, Brasília, DF, Brasil. Disponível em: https://www.camara.leg.br/. Acesso em: 1 out. 2025.

BRASIL. 2024. Versão final do Substitutivo ao Projeto de Lei nº 2.338/2023: propostas da Autoridade Nacional de Proteção de Dados (ANPD). Câmara dos Deputados, Comissão Especial sobre Inteligência Artificial no Brasil, Brasília, DF, Brasil. Disponível em: https://www.gov.br/anpd/pt-br/assuntos/noticias/anpd-apresenta-propostas-de-alteracao-do-substitutivo-ao-pl-2338-sobre-inteligencia-artificial/Versao_final_07.05.24_PL_2338_Substitutivo_Comissao_Especial_Propostas_ANPD1.pdf. Acesso em: 1 out. 2025.

FERNANDES, A. A.; COSTA, I.; ABREU, V. F. 2025. Governança da inteligência artificial: estrutura, desafios e práticas para organizações inovadoras. Brasport, São Paulo, SP, Brasil. E-book. Disponível em: https://plataforma.bvirtual.com.br. Acesso em: 20 set. 2025.

GABRIEL, M. 2025. A (r)evolução das habilidades para o futuro do trabalho na era da inteligência artificial. Editora Senac São Paulo, São Paulo, SP, Brasil.

GIL, A. C. 2008. Métodos e técnicas de pesquisa social. 6. ed. Atlas, São Paulo, SP, Brasil.

Koene, A.; Clifton, C.; Hatada, Y.; Webb, H.; Patel, M.; Machado, C.; LaViolette, J.; Richardson, R.; Reisman, D. 2019. A governance framework for algorithmic accountability and transparency. Disponível em: https://doi.org/10.2861/59990. Acesso em: 2 out. 2025.

SILVA, A. P.; COELHO, A. Z.; FEFERBAUM, M.; SILVEIRA, A. C. R. D. 2023. Ética, governança e inteligência artificial. Almedina, São Paulo, SP, Brasil.

Article originating from the Final Course Work of the Specialization in Compliance and ESG of the MBA USP/Esalq

To learn more about the course, click here and access the MBX Academy platform