Marianna Rozumna

Prof. Paulius Pakutinkas

How the era of AI change compliance

1. Introduction: general problem overview and aim of the research

We live in the age of digitalization and digital technologies that is very much developing
artificial intelligence and machine learning, which in turn affects the continuous development
of technologies and products. Artificial Intelligence is fundamentally changing the approach of
compliance in companies. With the help of machine learning algorithms and natural language
processing, artificial intelligence systems, companies can simplify processes in which large
amounts of data are analyzed, patterns are identified, and similar or different factors are
identified from certain standard algorithms or procedures. The impact of Artificial Intelligence
on compliance will be particularly significant in areas such as monitoring, risk assessment and
fraud detection. Artificial Intelligence will help compliance teams automate manual tasks,
reduce human error, and optimize resource allocation.
At the beginning of 2024, an important trend shaping the future of Artificial Intelligence
is Artificial Intelligence legislation. The rapid development of artificial intelligence
technologies has not only captured our imagination, but also attracted the scrutiny of global
policymakers. Countries around the world, including the United States and the European Union,
are on the verge of introducing revolutionary Artificial Intelligence regulations. At the same
time, the adoption of the long-awaited legislation and active discussion of Artificial Intelligence
regulation does not mean that it was not used in the compliance area before the adoption of the
regulation. Active usage of the Artificial Intelligence (AI) tools has been announcing in new
releases by Serious Fraud Office on April 10, 2018 stated that: “Today, the Serious Fraud Office
announced a significant upgrade to its document analysis capability as artificial intelligence is
made available to all of its new casework from this month.
By automating document analysis, AI technology allows the SFO to investigate more
quickly, reduce costs and achieve a lower error rate than through the work of human lawyers,
alone.
Able to process more than half a million documents a day, a pilot “robot” was recently
used to scan for legal professional privilege content in the SFO’s Rolls-Royce case at speeds
2,000 times faster than a human lawyer. Building on this success, ‘Axcelerate’ a new AI
powered document review system from OpenText, is now being rolled out alongside the robot,
and will enable SFO case teams to better target their work and time in other aspects of
investigative and prosecutorial work…” (Serious: Serious Fraud Office. (2018, April 10). AI
Powered ‘Robo-Lawyer’ helps step up the SFO’s fight against economic crime. Retrieved
October 2019.)
Companies will always have a need to implement both software products that improve
compliance with the necessary regulatory requirements and integrate artificial intelligence into
their own systems to obtain an effective compliance system. The use of artificial intelligence
and machine learning in companies significantly improves the performance of automated
repetitive tasks, in turn, allocating time for strategic issues of solving compliance problems. An
important part of the integration of artificial intelligence into the company's systems can be
used to identify risks and implement risk management processes, build certain algorithms to
predict consequences, analyze current transactions and decisions with the calculation of
financial losses, which allows companies to identify potential inconsistencies that may affect
the stability of the company.
Thus, it is obvious that the use of artificial intelligence has been going on for quite some
time; artificial intelligence is used for decision-making by both authorized state bodies and
companies, and therefore can have both negative and positive impact on the activities of any
company. That is why in this research the role of artificial intelligence in compliance will be
explored both in terms of positive and potentially negative impact on the company's operations’
activity. The goal is to foster responsible and balanced development of the AI and how the
development of the AI influence on the change of compliance risk assessment in companies.

2. Objectives
The potential impact of artificial intelligence in all sectors of the economy is critical to
affecting the operations of any company, even if it does not use artificial intelligence or
language models in its own operations. This requires determining what risks companies expect
in the business environment in connection with the active development of artificial intelligence
and its further use by any companies or individuals. The integration of artificial intelligence
into business operations is a game-changer for companies, driving them toward innovative
change. While artificial intelligence offers significant opportunities for efficiency gains, it also
creates a number of risks for companies.
Many companies are implementing an AI audit policy to understand the specific risks to
the AI tool and identify options to mitigate the risks. Understanding and managing compliance
risks associated with AI can be challenging, as assessing such risks must include the ability to
anticipate or account for developments in AI technology and the use of a sufficient database.
 The essay aims to investigate two issues regarding the development and impact of
artificial intelligence on the company's compliance system, which is to study the external
factors caused by the development of AI on risk assessment and approaches to risk assessment
within the company depending on such external factors; as well as the use of AI in determining
such risks and building a reliable automated system for assessing the risks of the company's
operating activities when compiling a compliance risk matrix.
The essay aims to explore two issues: the first one relates to the development and impact
of artificial intelligence on a company's compliance system which is to examine the external
factors that are caused by the development of AI on risk assessment and approaches to risk
assessment within a company depending on such external factors; from other side - the use of
AI in identifying such risks and building a reliable automated risk assessment system for a
company's operations in compiling compliance risk assessments and risk audits.
The essay explores how the development of new language models and the development
of artificial intelligence affects the stability of a company's operations and plays a key role in
building a compliance risk management system in any industry. AI is a valuable tool for risk
management through the use of advanced algorithms and data analysis to improve risk
assessment and mitigation strategies. Whether or not it is applied in many industries, the
contribution of AI is changing the way we assess potential risks and address them. Thus, the
positive impact and prospects for further development of technologies in the qualitative
construction of business processes and business operational decision-making are being
investigated. At the same time, the essay touches upon the topic of increasing and, in fact,
creating additional risks that have arisen due to the development of artificial intelligence and
exist regardless of whether the affected company implements artificial intelligence in its
business processes or activities.
These impacts and factors will be explored without focusing on legislative regulation,
with an understanding of the basic requirements of European artificial intelligence legislation
(Serious: European Parliament. (n.d.). EU AI Act. Retrieved from
https://www.europarl.europa.eu/doceo/document/TA-9-2024-0138-FNL-COR01_EN.pdf on
April 29, 2024, at 10:30 AM.) and the existence of separate legislative regulation in other
countries. The purpose is to research, evaluate, and disclose insights into the requirements of
artificial intelligence regulation and the basic requirements for building a compliance system,
standards for building compliance systems, security, risk management, and internal audit,
revealing the opportunities and issues arising from the use and development of artificial
intelligence and language models, regardless of the existence and development of regulatory
legislation for their technological development.
 The research involves identifying gaps and ambiguities, particularly in the context of
safety, liability, products, intellectual property, data protection, etc. The impact of the
development of artificial intelligence on the development of not only the tech industry, but any
industry is impressive, especially with regard to security issues and fairness or reasonableness
in decision-making. The essay will critically assess the current situation with negative attitudes
or actual ignoring of the development of artificial intelligence by companies, or the application
of artificial intelligence in their own activities without implementing or developing appropriate
necessary processes to engage in risk assessment related to artificial intelligence and large
language models. This research also aims to identify how artificial intelligence can change risk
assessment, focusing on its application to build a comprehensive risk assessment and improve
risk assessment mechanisms.
This analysis can serve as a basis for discussions within business associations or
companies on improving the risk management system and rethinking the risk assessment system
in building a sustainable business.

Methods
To achieve the comprehensive analysis of the discussed objectives, a diverse range of
research methods will be employed, each offering a unique window into this complex
landscape.
An extensive literature review of existing academic research, articles, and reports
pertaining to artificial intelligence, technology, risk management, data protection, and ethical
considerations will be conducted. A critical component of this research involves evaluating the
existing frameworks and consequences of developing of the AI in compliance risk assessment
system. This will include a general overview of relevant legislation and methods of AI
appliance in day-to-day compliance system of the company that can generate, evaluate or
influence on possible risks in business operation system.
Qualitative data, including statistical data, quantitative analysis, and market research, will
be sourced from open surveys and market research results. This qualitative data will be analyzed
to identify trends, emerging issues, and public sentiments related to the use of AI in compliance
and risk management. These methods will provide a comprehensive understanding of the
current state of AI technologies in risk assessment and the legal and ethical challenges they
present.
In addition, the research will include interviews with experts and panel discussions to
gain a deeper understanding of the practical challenges and opportunities that AI presents in the
compliance space. This approach will provide a more detailed understanding of how AI
technologies are currently being implemented and their impact on legal and ethical frameworks
in different jurisdictions.
Together, these methods will provide a comprehensive understanding of the current state
of artificial intelligence technologies in compliance risk assessment and the complex legal and
ethical issues they pose. The research aims to offer actionable recommendations for
policymakers, lawyers, and industry leaders to effectively address these challenges.

3. Findings
Compliance programs face increasing challenges as regulations become stricter, sanctions
become more severe, and threat landscapes rapidly evolve. To effectively ensure compliance,
compliance leaders need a proactive and strategic view of the risks in their organizations. While
basic risk assessment techniques provide a basic capability, it is now necessary to utilize
advanced techniques to provide the accuracy and foresight required in an environment of
greater uncertainty and higher stakes.
As artificial intelligence has become an important driver of innovation across industries,
its application to improve risk management processes is both promising and risky. Risk
assessment, an important component of any compliance program and sustainable business, has
traditionally relied on non-automated processes to assess, identify, analyze and mitigate risks.
However, the integration of artificial intelligence into this process is having an impact on the
effectiveness of the risk assessment process and system. Artificial intelligence can be used for
risk assessment by transforming traditional risk management approaches. The integration of
artificial intelligence into risk assessment paves the way for better risk identification, analysis,
and management in a variety of contexts. While the efficiency and complexity of AI-driven
processes are notable advantages, challenges related to data dependency, complexity, and the
need for human oversight remain. As artificial intelligence technology continues to evolve, it is
expected that its application in risk management will become more sophisticated.
Artificial intelligence systems are adept at collecting and processing large amounts of
data from a variety of sources, including historical and system data. These systems then analyze
the data thoroughly, identifying patterns, trends, and anomalies that may indicate potential risks.
According to the Moody’s report “Navigating the AI landscape INSIGHTS FROM
COMPLIANCE AND RISK MANAGEMENT LEADERS” it is shown the data shown the data
on what types of sphere companies use AI based on the conducted survey (Serious: Moody’s.
(n.d.). Navigating the AI landscape report (p. 10). Retrieved from
https://www.moodys.com/web/en/us/site-assets/ma-kyc-navigating-the-ai-landscape-
report.pdf on April 29, 2024, at 10:30 AM.)
 “PUTTING AI TO WORK Understanding where AI is being implemented today is key
to seeing how it will impact risk and compliance tomorrow. Unsurprisingly, as teams are forced
to deal with ever-increasing levels of information, 63% of companies actively using or in a trial
phase with AI are using it for data analysis and interpretation.”

Meeting and fulfilling the preceding requirements that must be considered and
implemented before implementing AI in compliance risk and operational efficiency programs
is an important and mandatory prerequisite for unlocking the potential of AI. This strategic
approach not only aligns with technological advances, but also supports business sustainability
goals in long-term risk management. (Serious: United Nations Global Compact. (n.d.).
Sustainable goals. Retrieved from https://unglobalcompact.org/sdgs/about on April 29, 2024, at
10:30 AM.)
Integrating artificial intelligence into risk management has the potential to transform
traditional practices by providing more accurate risk assessments and proactive risk mitigation
strategies. To realize these benefits, companies must create a fundamental framework that
supports sophisticated AI technologies.
Key prerequisites for effective integration of AI into risk management
High-quality data
The effectiveness of AI models depends heavily on the quality of the data used for
training. High-quality, relevant, and unbiased data is paramount to ensuring the accuracy and
reliability of models.
The dependence of AI systems on large, high-quality data underscores the need for strict
data governance protocols to ensure AI is effective in applications such as compliance risk
management. Companies can implement few types of strategies to address data availability,
reliability, algorithmic complexity, and ethical use of AI:
1. Data collection: Develop comprehensive data collection strategies that prioritize
diversity and scope to cover the various scenarios in which an AI system may operate.
2. Data Quality Assurance: Implement robust quality checks to clean and validate data to
ensure its accuracy and relevance to the problem at hand.
3. Data annotation: Use thorough data labeling processes, often with human supervision,
to ensure that training data accurately reflects real-world scenarios that the AI will encounter.
Risk management expertise
Risk management professionals play a crucial role in interpreting the results of artificial
intelligence and integrating this knowledge into strategic decision-making.
Compliance with regulatory requirements by establishing governance and ethical
principles
AI programs must comply with existing regulatory frameworks to ensure transparency
and accountability. All regulatory challenges associated with integrating AI, focusing on
compliance, transparency, and the ability to explain AI solutions should be met.
1. An AI ethical framework: Develop and implement ethical principles that govern the
use of AI, addressing issues such as fairness, privacy, and accountability.
2. Regulatory Compliance: Stay aware of and comply with local and international
regulations governing data use and AI deployment, ensuring legal compliance.
3. Stakeholder engagement: Involve diverse stakeholders, including legal, compliance,
and ethics boards, in the development and deployment of AI systems to ensure a broader
perspective and ethical standards are met.
SaaS for AI
Artificial intelligence systems require significant data processing capabilities, cloud-
based solutions have to be considerate to handle specific AI workloads.
AI engineer
Developing and maintaining artificial intelligence models requires professionals with
experience with artificial intelligence algorithms, data analysis, and programming
1. Advanced algorithm training: Use techniques such as feature engineering to improve
the interpretability of AI models without compromising their intended power.
2. Model Transparency: Use model agnostic techniques to explain AI decisions, such as
LIME (Local Interpreted Model Independent Explanations) or SHAP (SHapley's Supplemental
Explanations), which can help clarify how AI models reach their conclusions. (Serious: The
FinTech Times. (2022, January). How XAI Looks to Overcome AI’s Biggest Challenges.
Retrieved from https://thefintechtimes.com/how-xai-looks-to-overcome-ais-biggest-
challenges/ on April 29, 2024, at 10:30 AM.)
3. Ongoing monitoring: Establish ongoing monitoring of AI systems to quickly identify
and address any deviations or unexpected results, ensuring reliability and accuracy.
The effective implementation of AI in compliance risk assessment depends not only on
technological capabilities, but also on the ethical and responsible use of technology. By
establishing clear governance frameworks and adhering to ethical principles, companies can
mitigate AI-related risks such as bias, lack of transparency, and accountability issues. This
proactive approach ensures that AI technologies are used responsibly and sustainably,
contributing positively to companies’ goals and broader societal norms.
AI can be integrated into compliance frameworks to optimize processes and ensure
compliance in different business operation processes:
Contract analysis and creation system (Legal design)
1. Contract analysis: AI systems can be trained to understand legal language and
contractual obligations, allowing them to quickly review documents. These systems can
identify key provisions, obligations, and potential liabilities.
2. Risk Identification: Using machine learning algorithms, artificial intelligence can
analyze patterns and inconsistencies that may pose risks, flagging them for human review.
3. Suggestions for changes: AI can suggest necessary changes by comparing contracts to
a database of standard contracts.
Transaction monitoring
1. Real-time monitoring: AI can monitor financial transactions in real time, using the
analysis of standard transactions as a basis and rejecting non-compliant transactions for manual
review.
2. Suspicious Activity Detection: Using predictive analytics, AI models can examine
historical data to identify anomalies and potentially suspicious transactions or potential
fraudulent transactions.
3. Alert generation: AI systems can automatically generate alerts for suspicious
transactions, enabling quick and effective responses from system teams (e.g., blocking
transactions).
Intelligent risk compliance training using AI-powered tools such as virtual assistants and
chatbots
1. Individualized learning paths: AI can analyze an individual's learning pace, style, and
understanding to create customized training for that particular purpose. This way, each
employee learns what they need most and will be most effective for them.
 2. Real-time feedback and support: chatbots and virtual assistants can provide instant
answers to compliance-related questions, making learning much more dynamic and interactive.
This instant feedback can be a boost to learning and, in some circumstances, even correct
misunderstandings on the spot.
3. Gamification of learning: AI can incorporate gamification elements into compliance
training, making it more effective in driving employee engagement and motivation. For
example, you can get badges after completing modules or learn through quizzes to know where
the process is.
4. Continuous training and updates. AI systems will constantly update training material
in accordance with the latest regulations and practices. This ensures that the learning content is
up-to-date and thus can be maintained, thus gaining importance in areas where the legal
environment is changing at a very rapid pace.
5. Monitoring and reporting: AI can monitor through analytical reports each employee's
performance and completion rates in areas of learning where the employee is experiencing
difficulties. This data is invaluable in terms of responsible employees, which can be used to
identify risk areas and intervene if necessary.
6. Scenario-based training: With the help of artificial intelligence, training is designed for
interactive scenarios based on real-life scenarios, through which employees can rehearse their
lines in cases of complex compliance issues in a controlled environment.
Efficient resource allocation
1. Resource optimization. Automate routine compliance checks to reduce the use of
human resources.
2. Identify risk zones and send the right type of additional inspection.
3. Control of legislative changes. Artificial intelligence can be used to monitor current
legislative changes.
Automation of compliance reports
1. Simplifying complex tasks
2. Quickly extract the necessary data from large data sets, apply the appropriate regulatory
frameworks.
3. Reduces the need for significant manual labor and time, reduces the likelihood of
human error, and ensures that reports are compliant with applicable regulations.
Predictive insights and scenario modeling
1. AI's ability to process large amounts of historical data allows it to predict future risks.
2. Allow companies to prepare or change their strategies in advance.
 Integrating artificial intelligence into compliance processes provides numerous benefits,
from operational efficiency and increased accuracy to better compliance risk management and
regulatory compliance.
Compliance risk management is a critical aspect of business sustainability. AI algorithms
can analyze historical data, identify patterns, detect anomalies, and predict future risks with
high accuracy. This will allow companies to proactively identify and prevent compliance
operational risks before they occur. Here you can find one of the example on how trained AI
model identified certain type of risks based on the request:

ChatGPT generated risk assessment (Serious: Gomez, F. (n.d.). AI as a Risk officer from
https://medium.com/@francesca_52463/ai-as-a-risk-officer-how-chatgpt-fared-with-a-risk-
assessment-f73785bd9ef on April 29, 2024, at 10:30 AM)
Quantify compliance with AI tools (case) (Serious: Gomez, F. (n.d.). AI as a Risk officer
from https://medium.com/@francesca_52463/ai-as-a-risk-officer-how-chatgpt-fared-with-a-
risk-assessment-f73785bd9ef on April 29, 2024, at 10:30 AM)

Despite the positive aspects of using AI, there are also negative consequences of its use.
the following are the positive and negative aspects of using artificial intelligence in compliance
risk management.
The following criteria are positive in using AI in compliance risk assessment:
More accurate and timely execution of tasks:
AI systems use sophisticated algorithms to process a large set of data; they identify subtle
patterns and correlations that may be missed due to human error. Thus, this capability can
significantly improve the accuracy of risk assessment, allowing for more accurate prediction of
possible risk management strategies.
Increased efficiency:
Artificial intelligence will quickly and accurately perform large volumes of repetitive
tasks, such as data collection, processing, and preliminary analysis, which increases overall risk
management productivity.
Speed:
AI technologies support instant analysis and response to new risks. They can process data
streams in real time and offer instant solutions. The speed of processing and analysis ensures
that companies are able to respond quickly and efficiently to any threats or changes in the risk
environment.
Reduced costs:
AI can help automate routine, frequent tasks and thus reduce labor costs and human error.
 Scalability:
They can be designed to scale from a small business to the largest corporation or exactly
according to the size of the company. They can handle more data and risk models that are more
complex without the proportional growth in human resources and infrastructure.
Negative aspects of using AI in compliance risk management:
Low transparency:
The complexity of the algorithms in AI models creates "black boxes" that prevent
transparency of the decisions made. This lack of transparency makes it even more difficult for
risk managers to understand how these conclusions are reached, making it difficult to further
validate and trust AI assessments.
Bias:
AI models work according to the quality of the training data; if it is biased or
unrepresentative, the results will not be different. This will reinforce existing biases or create
additional risks of creating new, potentially unfair or harmful biases.
Expertise:
Developing and maintaining complex AI systems: In today's human and business
environment, maintaining various types of complex AI systems, such as data science and
machine learning, requires specialized knowledge. Lack of knowledge in the field may be one
of the main challenges, especially for smaller organizations, or access to technical education
may be one of the challenges in terms of the region of residence.
Privacy:
AI systems may need access to huge amounts of data, most of which is private personal
or commercial information. Protecting and securing this data to avoid intrusions and guarantee
privacy laws is a huge challenge (Serious: International Association of Privacy Professionals.
(n.d.). Privacy AI tracker. Retrieved from
https://iapp.org/media/pdf/resource_center/global_ai_law_policy_tracker.pdf on April 29,
2024, at 10:30 AM)
Legal and ethical considerations:
The convergence of AI in risk management can create complex legal and ethical issues.
These include liability for decisions made by AI, the possibility of creating a discriminatory
outcome, and compliance with ethical standards for AI operation. These concerns require
careful governance and a robust ethical framework to ensure responsible use of AI.
There are strategies to mitigate the disadvantages of using AI in risk management.
Increase transparency:
 Since AI systems can be opaque, increasing transparency is crucial. The use of
understandable artificial intelligence (XAI) techniques allows stakeholders to understand and
trust AI solutions by explaining how models process inputs to produce outputs.
Address bias:
AI biases can distort risk assessment and lead to unfair results. It is important to use
diverse and representative datasets to train AI models. Methods such as bias removal algorithms
can adjust the data or model to minimize or eliminate these biases, providing fairer and more
accurate results.
Technical expertise:
Lack of technical know-how can be a significant barrier to effective use of AI. Companies
may consider creating special training programs for their staff or partnering with AI consulting
firms. This approach helps to build internal expertise and ensures that the organization can keep
up with AI advances.
Enhance security:
AI systems are susceptible to various security risks, including data breaches and
cyberattacks. Implementing advanced security measures, such as encryption, multi-factor
authentication, and regular security audits, can protect sensitive information and safeguard AI
systems from external threats.
Legal and ethical issues:
Compliance with legal standards and ethical norms is of paramount importance.
Organizations should develop comprehensive policies that define the scope and limitations of
AI use. These policies should ensure ethical standards are upheld, ensure legal compliance, and
promote accountability. It is also important to have mechanisms in place to monitor and enforce
these guidelines.
Holistic approach:
Integrating these strategies requires a holistic approach that takes into account technical,
ethical, legal and security aspects. By addressing these areas together, organizations can
leverage the benefits of AI while minimizing the associated risks.
“At the same time there is a big negative influence of AI development on companies
irrespective of whether they use AI or not and which has to be evaluated in course of the process
of the compliance risk assessment. In general, there are a lot of cyberthreats that influence on
day-to-day activity of the company. Here we can discuss the few of the discusses threats to
show the global development and influence of the AI on business activity. As an example, we
can overview the Sumsub’s statistics show that the prevalence of deepfakes fraud grew
considerably from 2022 to 2023:
 From 2022 to 2023, the proportion of deepfakes among all fraud types increased by
4,500% in Canada, by 1,200% in the U.S., by 407% in Germany, and by 392% in the UK.
In Q1 2023, the most deepfakes came from Great Britain and Spain with 11.8% and 11.2%
of global deepfake fraud respectively, followed by Germany (6.7%), and the Netherlands
(4.7%). The U.S. held 5th place, representing 4.3% of global deepfake fraud cases
Last quarter, a high proportion of fraud consisted of deepfakes in Australia (5.3%),
Argentina (5.1%), and China (4.9%).
To put this in absolute numbers, from 2022 to 2023, the proportion of fraud consisting of
deepfakes increased from 0.1% to 4.6% in Canada, from 0.2% to 2.6% in the U.S., from 1.5%
to 7.6% in Germany, and from 1.2% to 5.9% in the UK.” (Serious: Sumsub. (n.d.). Learn how
fraudsters can bypass your facial biometrics. Retrieved from https://sumsub.com/blog/learn-
how-fraudsters-can-bypass-your-facial-biometrics/ on April 29, 2024, at 10:30 AM.)
Phishing: Phishing involves the use of fake links, emails, and websites to gain access to
a consumer's sensitive information - usually by installing malware on the target system. This
data is then used to steal other people's identities, gain access to valuable assets, and overload
email inboxes with spam. During elections, phishing emails can be disguised as donation emails
that encourage a citizen to click a link thinking they are donating to a candidate, but are actually
playing a bad actor's game.
Robotic calls, impersonations, and AI-generated voice or chatbots: As seen in New
Hampshire when an automated caller impersonated President Biden urging citizens not to vote,
the election season will see an increase in impersonations of pollsters or political candidates to
falsely gain trust and obtain sensitive information.
Deepfakes: With the advancement of artificial intelligence, deepfakes have become
incredibly realistic today and can be used to imitate your boss or even your favorite celebrity.
Deepfakes are videos or images that use artificial intelligence to replace faces or manipulate
facial expressions or speech. Many of the deepfakes we encounter on a daily basis will be in
the form of a fake video showing a person saying or doing something they may never have
done. This is expected to be especially prevalent this election season with the risk of diplomats
creating fake videos to impersonate candidates. Even outside of the US, such as in the UK, there
are concerns that diplomats could be used to rig elections.
The popularization of generative artificial intelligence is a new twist, as cybercriminals
can use these systems to improve their attacks:
Create better-written messages to trick recipients into providing sensitive data, clicking a
link, or downloading a document.
 Simulate the look and feel of emails and corporate websites with a high degree of realism,
avoiding arousing suspicion in victims.
Clone people's voices and faces and create deep voice or image fakes that attacked people
cannot detect. An issue that can have a significant impact on fraud, such as CEO fraud.
Respond effectively to victims thanks to the fact that generative AI can now carry on
conversations.
Launch social engineering campaigns in less time by investing fewer resources and
making them more complex and difficult to detect. Generative AI that is already on the market
can write texts, clone voices, or create images and program websites.
In addition, generative artificial intelligence opens the possibility of increasing the
number of potential attackers, as by allowing them to perform numerous actions, they can be
used by criminals without the necessary resources and knowledge.
As artificial intelligence continues to play an important role in fraud, it is crucial for
individuals and organizations to remain vigilant Thus, it is obvious that the standard processes
developed to identify and assess risk compensation are currently not sufficient to determine the
level of threat and the likelihood of a particular risk occurring, as the development of AI allows
artificial intelligence to bypass algorithms and processes that are or have been standard for a
long time or, including those developed for standard processes using AI.(Serious: Toh, A. (2018,
June 1). Are You Still Watching/ How Netflix Uses AI to Find Your Next Binge Worthy Show.
Nvidia. Retrieved October 2019.)

4. Conclusion:
In conclusion, AI integrated into the assessment of compliance risk represents a really
transformational shift in the use of regulatory practice. The ability of AI to process and analyze
huge datasets rapidly with accuracy substantially improves efficiency and effectiveness in the
identification and mitigation of potential compliance risks. However, it comes with its own set
of major cybersecurity challenges. As more artificial intelligence systems are becoming relied
upon, great concern is being pushed for very strong cybersecurity protection of data from
breaches and protection from more sophisticated cyber attacks.
In addition, AI deployed in compliance must tread carefully on issues surrounding data
privacy and the peril for algorithmic bias. This is a major concern in that the security of data
must be maintained while developing unbiased, equitable AI systems. It means those systems
need constant auditing to ensure they measure up to changing standards of morality and
regulation.
 Those are the cybersecurity and ethical challenges that organizations will have to rise in
a manner to fully exploit the potential that AI will bring in improving assessment of compliance
risk. This can include strict data protection policies, clear AI operation protocols, and relentless
training of compliance personnel in AI and cybersecurity standards. It is the responsibility of
regulatory bodies to ensure that the guidelines for the responsible and effective use of such tools
by keeping pace with technological advancement are updated and strictly enforced.
Looking forward, AI will very much redefine compliance practices together with
cybersecurity and human expertise. This approach not only encourages more resilience and
adaptability in compliance strategies but also unmovable adherence to ethical and legal
standards. The sensitive infusion of AI into compliance frameworks is required to develop a
secure, just, and compliant regulatory environment in the digital age.

Key Recommendations:
The compliance field in AI risk assessment pertains to various key areas for managing
effectively the potential risks linked with the deployment and use of AI systems. Here are some
of the recommendations:
1. Transparent and Explainable AI Models: Ensure that the AI systems used in the course
of compliance are transparent and, hence, can afford explainable decisions. This will help in
understanding the way decisions are made and, hence, uncover biases within the system.
2. Regular Audits and Updates: Continue running AI system audits regularly to keep the
operation in line with its intended operation and regulatory requirements. This would include
updating the AI algorithms and models in line with the new set of regulations and compliance
requirements.
3. Data Privacy and Security: Strong data governance practices need to be followed to
safeguard the sensitive information accessed or generated by the AI systems. Adhere to various
data protection regulatory compliances, such as GDPR or HIPAA, based on the geography and
industry.
4. Bias Detection and Mitigation: Develop and implement methodologies for detecting
and mitigating biases in AI models. This is key in ensuring that, if fairness is not achieved, the
decision could always be discriminating.
5. Engage stakeholders: Engage across different stakeholders, which include the legal,
compliance, and IT teams, to seek their advice in alignment with AI strategies, business
objectives, and already-existing compliance requirements.
 6. Ethical AI Framework: Develop an AI framework of ethics conforming to
organizational value systems and compliance imperatives. This framework will set the
guideline for developing and implementation of AI Technologies.
7. Training and Awareness: Train employees in the context of AI implications for
compliance, including the capabilities and limitations of AI technologies.
8. Compliance-specific AI Development: Develop AI tailor-made to offer solutions to the
problem of compliance. This might include developing customized solutions or modifying
existing technologies to fit the need for regulatory requirements better.
9. Impact Assessment: Carry out thorough impact assessments prior to deploying AI
systems, understanding how the use of AI may impact the compliance status. It should include
scenario planning and potential risk identification.
10. Work with Regulators: Interact with the regulators to ensure that the AI systems being
put in place meet required compliance standards and, more so, take measures in advance to
manage any regulatory changes that may come into effect.
These strategies will help to manage the risks in compliance better while ensuring that
they continue to operate within the regulatory adherence and ethics of the business.

Serious:
1. European Central Bank. (2021). Seventh Report on Card Fraud. European Central Bank.
Retrieved from https://www.ecb.europa.eu/pub/cardfraud/html/ecb.cardfraudr
eport202110~cac4c418e8.en.html on April 29, 2024, at 10:30 AM.
2. European Parliament. (n.d.). EU AI Act. Retrieved from
https://www.europarl.europa.eu/doceo/document/TA-9-2024-0138-FNL-COR01_EN.pdf
on April 29, 2024, at 10:30 AM.
3. Gomez, F. (n.d.). AI as a Risk officer.
4. International Association of Privacy Professionals. (n.d.). Privacy AI tracker. Retrieved
from https://iapp.org/media/pdf/resource_center/global_ai_law_policy_tracker.pdf on
April 29, 2024, at 10:30 AM.
5. Kayid, A. (2020). The role of Artificial Intelligence in future technology. Retrieved from
https://www.researchgate.net/publication/342106972_The_role_of_Artificial_Intelligence
_in_future_technology on April 29, 2024, at 10:30 AM.
6. Moody’s. (n.d.). Navigating the AI landscape report (p. 10). Retrieved from
https://www.moodys.com/web/en/us/site-assets/ma-kyc-navigating-the-ai-landscape-
report.pdf on April 29, 2024, at 10:30 AM.
7. Serious Fraud Office. (2018, April 10). AI Powered ‘Robo-Lawyer’ helps step up the SFO’s
fight against economic crime. Retrieved October 2019.
8. Sumsub. (n.d.). Learn how fraudsters can bypass your facial biometrics. Retrieved from
https://sumsub.com/blog/learn-how-fraudsters-can-bypass-your-facial-biometrics/ on April
29, 2024, at 10:30 AM.
9. The FinTech Times. (2022, January). How XAI Looks to Overcome AI’s Biggest
Challenges. Retrieved from https://thefintechtimes.com/how-xai-looks-to-overcome-ais-
biggest-challenges/ on April 29, 2024, at 10:30 AM.
10. Toh, A. (2018, June 1). Are You Still Watching/ How Netflix Uses AI to Find Your Next
Binge Worthy Show. Nvidia. Retrieved October 2019.
11. United Nations Global Compact. (n.d.). Sustainable goals. Retrieved from
https://unglobalcompact.org/sdgs/about on April 29, 2024, at 10:30 AM.