The communications department has identified several steps to ensure the national artificial intelligence (AI) policy comes to fruition, envisaging completing the process in the 2027/2028 financial year.
This came to light during the department’s briefing before Parliament’s Portfolio Committee on Communications and Digital Technologies on Tuesday.
The meeting sought to provide the portfolio committee with insights into the crafting and implementation of the national AI policy, including ethical AI guidelines for national adoption.
Mlindi Mashologu, deputy director-general (DDG): digital society and economy in the Department of Communications and Digital Technologies, detailed the steps the department is taking following the publishing of the draft AI policy in 2024.
Mashologu’s presentation emphasised the urgency of establishing a nationwide AI policy to leverage AI’s benefits, such as job creation and efficiency improvements, while mitigating risks like job losses through digital skills training.
He told MPs that all five government clusters have concurred that the policy must be taken to Cabinet for gazetting.
Resultantly, it will now undergo Cabinet approval and public consultation, with sector-specific strategies to follow, said the DDG. “What we’re looking toward is going through the different cluster committees [early March] …and then Cabinet later in March.
“From then onwards, the policy will be gazetted for public comments; we envisage 60 days for public comments. We will then start to refine the policy for it to be adopted and then finalise that part of the development process in the next financial year (2026/2027).”
Government’s financial year kicks-off on 1 April, signalling the start of the 2026/2027 financial year.
Mashologu’s presentation revealed that the nationwide AI policy, inclusive of regulations and guidelines, will be developed by 2027/2028.
International benchmarking
With AI infiltrating almost every industry, policymakers across the global community have increasingly focused on its regulation. This, to ensure AI is created responsibly and ethically from the start, to ensure citizens, businesses and governments are protected from some of the potential AI risks.
In the European Union (EU), standards for AI regulation and enforcement worldwide have been set up through the EU AI Act. The legislation is described as one that specifically regulates AI technologies and opens vast opportunities for innovation across various sectors globally.
Closer to home, African nations Nigeria, Mauritius and Rwanda are already on their way to developing their own AI strategies and policies.
South Africa has been somewhat on the back foot, with no national AI roadmap in place.
However, Mashologu stated that work on the AI framework started in the background in 2020, adding that SA led the development of an AI blueprint for the African Union in 2021.
The country contributed to the continental AI strategy and established AI hubs at various universities, he stated.
“The draft national AI policy framework was developed and published in August 2024. It benchmarked strategies from other countries, like the Netherlands, Chile, Thailand, Norway and Rwanda, as well as development that has been done in terms of the EU AI Act.”
He explained that the national AI policy is being developed to ensure the risks and benefits of AI are equally distributed across society.
“Currently, AI is distributed unevenly. We are saying that we need to level the playing ground with the factor that is called intergenerational equity. We’re also trying to position AI as a strategic [pillar] to support inclusive growth, societal development and strengthen the government capacity to regulate, adopt and scale AI responsibly.
“One area that I need to highlight is in terms of the regulation of AI; we see in some instances that this area is not going to be regulated by one entity. So, ICASA [the Independent Communications Authority of SA] is going to be part of the regulators…we also spoke to the likes of the Information Regulator, as well as all the other regulators, to see the areas they will be regulating.
“We need to make sure the policy is finalised but also ensure sectoral strategies can be developed. This includes the ethical guidelines, regulations to be developed, and then go through the final implementation of the policy and all its recommendations.”
He said there will be a need for a phased-approach in the adoption of AI, explaining that the policy won’t be a blanket framework for all sectors, but that different sectors will have to use it to craft sector-specific strategies that are tailor-made for their environments.
Pushed for time
Some MPs sounded the alarm over the timeline risks, lack of skills, data residency, regulatory oversight, the impact of AI on employment and whether there would be sufficient mechanisms to raise public awareness during the consultation process.
The discussion also touched on the need for inclusive AI data sets, the role of public consultations and the ethical considerations in AI applications, like crime detection.
Furthermore, MPs highlighted the importance of addressing systemic discrimination and ensuring digital inclusion, along with the need for a comprehensive job impact assessment and clear accountability mechanisms.
Communications minister Solly Malatsi previously told ITWeb that SA’s compliance and consultation approach to policy development is multi-layered, stressing that there has not been an intentional hold-up in finalising the AI policy.
On the timelines, some MPs yesterday wanted to know if the target could be moved earlier than the 2027/2028 financial year.
Mashologu responded: “We will try on our side to ensure the policy is brought forward. For us, if we can get the policy gazetted later in the coming month, we should be able to push through and ensure that by mid of the next financial year we will be finalising the final policy.”
Department director-general Nonkqubela Jordan-Dyani echoed the MPs’ sentiments that the department needs to review the time frames.
“A lot of work has already been done,” she stated. “I think it’s very important that we set forward a nationwide policy position with regards to where we want to take artificial intelligence and where we see the opportunities, in terms of the innovation and improving efficiencies.
“’We’re seeing there are more benefits in terms of the adoption of artificial intelligence than there would be in terms of the risk, and I think the most inherent risk is about the issue of job losses.
“We are trying to mitigate this on the digital skills level…we’re collaborating with the TVET colleges. We had a workshop yesterday to determine how we expand and accelerate that programme. For our digital skills, we will make sure we build the skillset identified by the industry.”
Share