The EU AI Act is Here: Essential Compliance Deadlines for Your Business

The AI Act’s phased implementation allows companies time to adjust, but the requirements are rigorous. High-risk AI systems, such as those used in healthcare, law enforcement, and critical infrastructure, will face the most stringent regulations. These include ensuring data quality, implementing anti-bias measures, and conducting comprehensive risk assessments. Non-compliance can result in hefty fines, up to €35 million or 7% of global annual revenue for serious breaches [1][2].

The European Union’s landmark Artificial Intelligence regulation, known as the EU AI Act, has been officially published in the EU Official Journal. This significant milestone sets the clock for the enforcement of new legal standards designed to regulate AI applications based on their risk levels. As of July 12, 2024, the AI Act is now formally published, and it will come into force on August 1, 2024, initiating transition periods that enable organizations to adapt to the new compliance regulations [1].

Key Compliance Deadlines

With the Act’s publication, various compliance deadlines are now approaching for technology companies operating within the EU. Here are the critical dates to be aware of:

August 1, 2024: The AI Act comes into force, starting the countdown for various compliance deadlines [2].

February 2, 2025: Initial compliance phase begins. Prohibited AI systems, such as those used for biometric categorization without consent, must be withdrawn from the market. Companies must also start complying with transparency obligations, ensuring users are aware when they are interacting with AI [2].

May 2, 2025: Developers must adhere to newly established codes of practice that detail benchmarks for compliance and performance indicators [2].

August 2, 2025: General-purpose AI models, including chatbots, must meet specific transparency and copyright compliance requirements [2].

August 2, 2026: Full compliance with the Act’s provisions is required for most AI applications. This includes the establishment of AI regulatory sandboxes to support innovation and compliance [1].

August 2, 2027: Deadline for high-risk AI systems to comply with stringent requirements regarding risk assessments and human oversight [1].

Implications for Technology Companies

The AI Act’s phased implementation allows companies time to adjust, but the requirements are rigorous. High-risk AI systems, such as those used in healthcare, law enforcement, and critical infrastructure, will face the most stringent regulations. These include ensuring data quality, implementing anti-bias measures, and conducting comprehensive risk assessments. Non-compliance can result in hefty fines, up to €35 million or 7% of global annual revenue for serious breaches [1][2].

Supporting Ethical AI Development

At Verdas AI, we are dedicated to empowering organizations to navigate these new regulatory landscapes while building ethical and high-quality AI solutions. To assist companies in meeting these stringent new regulations, we offer specialized training sessions, including the IEEE CertifAIEd™ Assessor Training. This program covers essential topics such as ethical transparency, privacy, accountability, and algorithmic bias, ensuring professionals are well-equipped to manage AI governance effectively [3].

Upcoming IEEE CertifAIEd™ Training Sessions

Verdas AI is offering upcoming training sessions to help professionals stay ahead in this rapidly evolving field. These sessions provide an opportunity to engage with leading experts in AI ethics and governance. The next training sessions are scheduled for July and August, and we encourage all professionals involved in AI development and deployment to participate.

Get More Info About Training

Staying informed and trained in AI ethics and governance is crucial for professionals navigating the complexities of AI technologies. By participating in these training sessions, professionals can ensure they are well-equipped to contribute to the responsible development and deployment of AI technologies, benefiting society as a whole.

For more information on upcoming training sessions, visit our training page [3].

References

[1] EU Official Journal – AI Act. Retrieved from: eur-lex.europa.eu

[2] European Parliament Press Release on AI Act. Retrieved from: europarl.europa.eu

[3] IEEE CertifAIEd™ Assessor Training. Retrieved from: verdas.ai

Welcome to the third entry in our blog series, “IEEE SA CertifAIEd™ Authorized Assessor Training: SIRIM Academy (Malaysia) – Participant Stories”.

In this series, we share inspiring stories and experiences of professionals who have recently undergone our IEEE CertifAIEd™ Authorized Assessor training. We believe that hearing directly from these professionals can provide valuable insights and motivation.

Below, you’ll find testimonials from a Deputy Director of Digital Demand Creation Department at MCMC, Malaysia.

Firdaus Azman: Can you introduce yourself? 

Mr. Ismail: I am Athari Ismail, the Deputy Director or Senior Executive at the Malaysian Communications and Multimedia Commission (MCMC) under the Digital Demand Creation department.

I oversee adoptions of artificial intelligence (AI), introduce programs and initiatives, and collaborate with external stakeholders. Especially in developing policies and governance frameworks on both national and international platforms.

I am currently helping the development of an AI Code of Ethics for the Communications and Multimedia industry. Additionally, I am actively involved in several high-profile committees and work groups focusing on AI ethics.

I am a permanent member for various high-level advisory body on AI convened by the Malaysian government and international bodies. Such as the Joint Coordination Action by the United Nations (UN) and International Telecommunication Union (ITU).

From 2022 till 2023, I contributed and took part in the development of an AI international governance for ASEAN.

Presently, I am an active member for an ISO work group – subcommittee of the Joint Technical Committee (JTC) for ISO. Where we are standardizing AI use cases and applications on behalf of Malaysia.

My various involvements (beyond his role as the Deputy Director) have given me broad perspectives on ethical challenges and opportunities in AI. Especially with my recent training on IEEE CertifAIEd™ Authorized Assessor programme. Where the training had further enriched my knowledge in this area.

Firdaus Azman: What have you gained from the training? 

Mr. Ismail: The IEEE CertifAIEd™ training had provided me with invaluable experience. To start, the training provided me with a comprehensive framework for assessing the ethical aspects of AI systems. Where I gained a deeper understanding through the lessons on criteria suites.

The interdisciplinary approach of the training (guided by the criteria suites) and training from AI Ethics experts have equipped me with the necessary tools to critically evaluate and ensure ethical deployment of AI technologies.

And as I have mentioned, my portfolios on creating AI policies and governance frameworks, benefitted from this training. I can see how the CertifAIEd™ programme aligns with upcoming regulation and positions me to stay ahead in the field of AI Ethics.
So I find the training particularly relevant to my current work at MCMC (and beyond), as we are developing the AI Code of Ethics for the Malaysian Communications and Multimedia industry.

The knowledge and framework provided by CertifAIEd™ are instrumental because it focuses on ethical governance. Aligning perfectly with our current objective of ensuring fairness, transparency, and accountability in AI developments.

Just to give you an example on how I have applied the concept learned through CertifAIEd™, I am now able to effectively assess ethicality of AI based products and systems, in terms of identifying potential biases and advocate for stronger privacy and security measures.

Additionally, the training has also enhanced my ability to engage stakeholders and guide industries towards ethical AI practice.

Firdaus Azman:  How do you find yourself applying the training in your work?

Mr. Ismail: One specific application of this training in my work is in the development of the AI Code of Ethics for the Communications and Multimedia industry.

It gave me a way of building potential frameworks and assessment tools that we may integrate into the code. Not directly but adapting what we have learned from CertifAIEd™ into the development of the code.

For example, the criteria suites can be used to evaluate existing AI applications for the industries. Where we may benchmark the current AI ethics practice with IEEE CertifAIEd™’s framework.

Additionally, the training insight into the suites (privacy and accountability) are helpful in helping us shape policies that protect user data and ensure responsible AI usage.

For a regulatory body overseeing the Communications and Multimedia industry, we have to look into different stakeholders’ interests. The training that we had allowed us to explore this, by looking at the interest of users and consumers.

The Concept of Operations (ConOps) allowed us to map and explore stakeholders according to an AIS context. And it also made us assess every dimension and spectrum of risk involved in the AIS context and ecosystem. 

Firdaus Azman: What are some of the IEEE CertifAIEd™ training benefits that you see for Malaysia and Southeast Asia?

Mr. Ismail: From my perspective, the vision that we should have for AI ethics in Malaysia and Southeast Asia is to establish a comprehensive and adaptable framework that can guide ethical development and deployment of AI across various sectors.

As you know, different sectors have different AI strategies, deployment, and also adoption. Which is one of the reasons that we are developing and introducing an AI Code of Ethics for Communications and Multimedia industry.

That is to pursue continuous collaborations with national and international bodies that aligns with our own standards and global best practices. Which is also a way for us to foster an environment where innovation thrives alongside ethical considerations.

These are also important factors when it comes to governing emerging technologies such as AI. In the Malaysian context, we are also looking at ways to ensure AI technologies contribute positively to economic growth and societal welfare.

I am personally committed to strengthening the Malaysian public awareness and understanding of AI Ethics, through which we would be able to promote trust and transparency in AI systems among users.

Lastly, the vision that I mentioned earlier can perhaps be realized by adopting frameworks such as IEEE CertifAIEd™, because of its ability to strengthen our envisioned ecosystem. 

Firdaus Azman: How do you go about pursuing public awareness?

Mr. Ismail: I would say this relates to my previous involvement in the Malaysian blockchain and AI committee. Where there was a need of us not being left behind in creating public awareness and ethical adoption. In response to this need, we introduced guidelines, but we did so in a careful manner to not stifle creativity and growth of the ecosystem.

In response to this question, we may perhaps explore ways of educating the public at an early level. Especially on how to be responsible in using AI technology in the Communications and Multimedia industry.

I am aware that there are a lot of challenges out there on AI, especially with the social media, deep fake, and tools to violate copyright or infringes copyright. Which may also be addressed by early education. For example, from the school up to university levels, incorporating some the ethos that are included in IEEE CertifAIEd™ framework.

About the Training:

The IEEE SA CertifAIEd™ Authorized Assessor Malaysia training program was conducted from May 13th to 17th, 2024. It involved 25 professionals from notable academic institutions, government agencies, and major private companies in Malaysia.

The program benefited from active collaboration between Malaysia’s SIRIM Academy, the Ministry of Science, Technology, and Innovation (MOSTI), the Institute of Electrical and Electronics Engineers Standards Association (IEEE SA), and Verdas AI.

United by a common interest in pursuing ethical AI practice and certification, these professionals spent four days in training followed by one day for the exam, which tested their knowledge of the IEEE CertifAIEd framework.

If you are interested in promoting ethical AI practices and becoming a certified assessor, visit our training website by clicking the button below for more information and to sign up for our next training session. Together, we can responsibly shape the development and integration of AI systems.

Welcome to the second entry in our blog series, “IEEE SA CertifAIEd™ Authorized Assessor Training: SIRIM Academy (Malaysia) – Participant Stories”.

In this series, we share inspiring stories and experiences of professionals who have recently undergone our IEEE CertifAIEd™ Authorized Assessor training. We believe that hearing directly from these professionals can provide valuable insights and motivation.

Below, you’ll find testimonials from the Head of Digital Demand Creation Department at MCMC, Malaysia.

Participant 2: Mrs. Jacinta Felix, Head of Digital Demand Creation Department, for Malaysian Communications and Multimedia Commission (MCMC)

Firdaus Azman: Can you introduce yourself?

Mrs. Felix: I am Jacinta Felix, the Head of Digital Demand Creation Department for the Malaysian Communications and Multimedia Commission (MCMC).

MCMC regulates the telecommunication, broadcasting, postal and courier industries. We also govern the Digital Signature Act.

I am currently responsible for three scopes of work: Artificial Intelligence (AI), 5G adoption, and emerging technologies. For the scope of AI, my team and I are looking into policy and strategic initiatives, potential regulation, and the adoption by the communications and multimedia industry. In addition, we are also working on 5G adoption by enterprises to spur Malaysia’s digital economy.

On emerging technologies, we are looking at how AI and other technologies are developing and interconnecting for the end users. Specifically, my team and I are exploring potential developmental work and regulatory developments revolving around AI and Emerging Technologies.

This effort is done with clarity of focus and structured balancing act of development and regulation. To ensure that each segment of focus does not stifle growth and innovation for the Malaysian industry.

Firdaus Azman: What have you gained from the training?

Mrs. Felix: I began my involvement in AI work for MCMC in January 2024, and obviously AI is a relatively new area for me.

I was expediting a self-learning mode for the first few months, absorbing everything and anything on AI that came my way. While on a steep learning curve, I have also made the conscious decision that all the acquired knowledge requires a structure to process these thoughts. Which would later ensure a sound policy and strategy deployment amidst rapid AI development.

This is particularly important, in which the policy and strategy implementation work would benefit from a structural process. Where it makes it so much easier to implement if everyone is on the same page in terms of understanding. Which is crucial for the success of any AI related work.

Hence, a certification course and training like the IEEE CertifAIEd™ Authorized Assessor training is just what we needed. And it came at a timely moment (as they are developing and deploying many policies and strategies on AI for Malaysia).

I am an example of a new entrant to AI, with an expedited learning curve in the first four months, and then came the IEEE CertifAIEd™ Authorized Assessor training that helped me to put things into perspectives. Particularly on what I have learned so far.

Though some would say the four days of training, with an exam at the end, are quite overwhelming. I believe that on the fifth and exam day, everything came together perfectly. Meaning, the lessons from the training made perfect sense at the end of the training (despite initial complexity and steep learning curve). Especially on the practical use cases and criteria suites (principles) that were carried out during the earlier four days.

Even as I am speaking to you now, the criteria suites came easily and naturally to my mind. It may have initially appeared extensive or complex, yet, upon completing the training course, one would find the four criteria suites (principles) easily remembered (perhaps due to its relevance or universality to the domain of AI ethics).

In my honest view, four days of training with a steep learning curve were initially overwhelming. However, it was worth sitting through the hours. Because at the end of the training course, I could still remember the four criteria suites and how they are applied in the IEEE CertifAIEd™ framework.

I would like to emphasize that it is not about the memorization of the suites, rather, the remembrance and comprehension of them. Specifically, how they can be applied in their line of work (particularly for AI professionals). For example, if I am to develop an AI system for the Commission, the training would help me to structure the whole development process according to the four criteria suites (principles).

Even in my current project of developing the AI Code of Ethics for the Malaysian Communications and Multimedia industry, I am able to sync or connect our current work with the four suites. So, there is definitely a high degree of relevance and usefulness with her line of work.

To conclude, I am confident that the four criteria suites are a good runway that helps us to put things into perspective and avoid being out of focus in our work related to AI governance.

I would like to take this chance to thank the Malaysian Ministry of Science, Technology, and Innovation (MOSTI), MCMC, and SIRIM Academy, for the training opportunity. I am humbled and grateful by the fact that I am now one of the pioneering members for the IEEE CertifAIEd™ Authorized Assessor program for Malaysia.

Moreover, I enjoy the exam experience because it does not test your memory. Rather, our understanding of the framework and criteria suites. In which, one must strategize and provide substantive justification for their answers.

Firdaus Azman:  How do you find yourself applying the training in your work?

Mrs. Felix: While some of the training programs that I had in the past would indirectly apply to my work, the IEEE CertifAIEd™ Authorized Assessor training program applies directly to my line of work.

This is from two perspectives; 1) policymaking and regulation development, and 2) AI adoption, for example how to adopt AI systems in a work environment.

In the first perspective, the criteria suites are helpful to me in structuring their policymaking and regulation development (as was deliberated earlier).

In the second perspective, because MCMC is also looking at adopting AI systems in our own organization, we would also later shape the Malaysian industries (telecommunication, broadcasting, postal and courier) on best principles to adopting AI systems (by referring to the criteria suites).

My team and I are very much structured and focused on building guardrails for both internal (MCMC) and external stakeholders (industries) on how to develop, deploy, and integrate AI technology.

MCMC as the regulator for Malaysia is the reference point for these industries in Malaysia for developing guidelines, regulations, and overall policy approaches in governing technologies.

In this context, we need to not only be able to catch up with the rapidly evolving AI technology and field, but also advise and provide timely guidance for industries in Malaysia. Referring to the IEEE CertifAIEd™ training, the criteria suites would help me in advising industries on how to develop, deploy, and integrate AIS in their activities.

Reflecting on my training experience and current application of my training, the criteria suites helped me in both policymaking and implementation. That is, on how to develop ethical AIS policies and implement them for Malaysia.

Firdaus Azman: What are some of the IEEE CertifAIEd™ training benefits that you see for Malaysia and Southeast Asia?

Mrs. Felix: MCMC is currently looking at sectoral governance of AI, specifically for the industries that we are responsible for. However, I strongly believe that it is also ideal for other industries to adopt the same framework (IEEE CertifAIEd™) via the criteria suites.

This is mainly due to the importance of standardization (harmonization) across all industries on the principles (criteria suites) to evaluate AI. Ensuring that all industries have common understanding and consensus on the critical aspect of AI which requires prioritization.

Having a standardised structure would allow a common language that translates well into integrating systems that are similar across industries, countries, and regions/continents.

The CertifAIEd™ framework is so structured that even though there is an emphasis on ‘subjectivity of AIS context’ as part of the understanding in CertifAIEd™ framework, the framework accommodates vastly different players (for example machine learning engineers and policy makers) to have a common platform and language to work upon in building responsible AIS.

I would add that the CertifAIEd™ framework has the potential to unify the currently different approaches across borders and regions (ASEAN to APAC). Because of the structure and standardized approach that everyone can tailor fit to their AIS locality and context. Which is an ideal outcome that I hope to have.

About the Training:

The IEEE SA CertifAIEd™ Authorized Assessor Malaysia training program was conducted from May 13th to 17th, 2024. It involved 25 professionals from notable academic institutions, government agencies, and major private companies in Malaysia.

The program benefited from active collaboration between Malaysia’s SIRIM Academy, the Ministry of Science, Technology, and Innovation (MOSTI), the Institute of Electrical and Electronics Engineers Standards Association (IEEE SA), and Verdas AI.

United by a common interest in pursuing ethical AI practice and certification, these professionals spent four days in training followed by one day for the exam, which tested their knowledge of the IEEE CertifAIEd framework.

If you are interested in promoting ethical AI practices and becoming a certified assessor, visit our training website by clicking the button below for more information and to sign up for our next training session. Together, we can responsibly shape the development and integration of AI systems.

Welcome to the first entry in our blog series, “IEEE SA CertifAIEd™ Authorized Assessor Training: SIRIM Academy (Malaysia) – Participant Stories”.

In this series, we share inspiring stories and experiences of professionals who have recently undergone our IEEE CertifAIEd™ Authorized Assessor training. We believe that hearing directly from these professionals can provide valuable insights and motivation.

Below, you’ll find testimonials from a Technology Consultant at SIRIM Academy, Malaysia.

Participant 1: Mr. Zulkefli, Technology Consultant at SIRIM Academy

Firdaus Azman: Can you introduce yourself?

Mr. Zulkefli: I am Zulkefli Mohd Dahlan, a Consultant for Technology Section at SIRIM Academy, under the Department of Quality and Technology.

I support one of SIRIM Academy’s current focus, that is on Training and Consultancy for Quality and ISO related activities, especially on emerging technologies.

Firdaus Azman: What have you gained from the training?

Mr. Zulkefli: Through the five days of training (including 1 day for exam), we have gained better understanding on ethical AI practice and AI ethics, IEEE CertifAIEd™ methodology, and how to apply and assess AI systems.

We have also learned fundamentals on Ethics Profiling (one of CertifAIEd™’s framework-processes), and how to conduct CertifAIEd™ assessment and certification processes. 

Firdaus Azman:  How do you find yourself applying the training in your work?

Mr. Zulkefli: It is a great opportunity for me to learn about the training (as a training consultant for SIRIM Academy for their future CertifAIEd™ training deliveries). Becoming an Authorized Assessor would allow me to conduct and facilitate our own future CertifAIEd™ training for SIRIM Academy.

The training also taught me collaboration process with IEEE SA and Verdas AI, and on how to conduct and deliver IEEE CertifAIEd™ training. I see great value in becoming both a trainer and an assessor for CertifAIEd™.

Additionally, because of my constant communication with academia and industry players involved in AIS development in Malaysia, the training has enabled me to understand and speak about AI Ethics and Governance.

Especially when talking about the four criteria suites; Ethical Algorithmic Bias, Ethical Transparency, Accountability, and Ethical Privacy. It is a good talking point to have with these players, especially with the recognition of an emerging AIS development and deployment field in Malaysia and Southeast Asia.

I would also like to encourage these players to take the CertifAIEd™ training, where I hope that this would better promote ethical AIS practice among them.

Firdaus Azman: What are some of the IEEE CertifAIEd™ training benefits that you see for Malaysia and Southeast Asia?

Mr. Zulkefli: At SIRIM Academy, we plan to encourage and nurture ethical AIS development in every aspect. Especially in upholding all four (if not most) of the criteria suites for Malaysia and Southeast Asia.

Currently, there are a lot of startup companies in Malaysia and Southeast Asia that are utilizing and embedding AIS into their business processes or systems. However, most of them are unaware or have little focus on AI ethics and assessment.

The CertifAIEd™ training had allowed me (and he believes this would also benefit Malaysia and Southeast Asia) with the knowledge on the required processes or steps for ensuring ethical AIS development, deployment, and integration.

Moreover, the emphasis of CertifAIEd™ framework on human values, such as ‘fairness’, and governance methods, would benefit those who seek to have certain structure in assessing and evaluating AIS. Hence, allowing for better control over AIS against unwanted or unintended output or outcomes.

About the Training:

The IEEE SA CertifAIEd™ Authorized Assessor Malaysia training program was conducted from May 13th to 17th, 2024. It involved 25 professionals from notable academic institutions, government agencies, and major private companies in Malaysia.

The program benefited from active collaboration between Malaysia’s SIRIM Academy, the Ministry of Science, Technology, and Innovation (MOSTI), the Institute of Electrical and Electronics Engineers Standards Association (IEEE SA), and Verdas AI.

United by a common interest in pursuing ethical AI practice and certification, these professionals spent four days in training followed by one day for the exam, which tested their knowledge of the IEEE CertifAIEd framework.

If you are interested in promoting ethical AI practices and becoming a certified assessor, visit our training website by clicking the button below for more information and to sign up for our next training session. Together, we can responsibly shape the development and integration of AI systems.

Exploring the Latest in
AI Governance and Regulation

References

[1] European Parliament. EU AI Act. Retrieved from: europarl.europa.eu
[2] IEEE CertifAIEd™ Training Overview. Retrieved from: verdas.ai

Exploring the Latest in
AI Governance and Regulation

Conclusion

Staying informed and trained in AI ethics and governance is essential for professionals navigating the complexities of AI technologies. The IEEE CertifAIEd™ Training offers valuable insights and skills needed to address the ethical and regulatory challenges posed by AI. By participating in these training sessions, professionals can ensure they are well-equipped to contribute to the responsible development and deployment of AI technologies, benefiting society as a whole.

References

[1] US Senate AI Roadmap Report. Retrieved from: schumer.senate.gov
[2] EDPS Guidelines on Generative AI Press Release. Retrieved from: edps.europa.eu
[3] EDPS Guidelines Document. Retrieved from: edps.europa.eu
[4] UK House of Commons AI Governance Report Summary. Retrieved from: publications.parliament.uk
[5] UK House of Commons AI Governance Report Full. Retrieved from: publications.parliament.uk
[6] IEEE CertifAIEd™ Training Overview. Retrieved from: verdas.ai