Artificial Intelligence and Data for Development in the Middle East and North Africa: Preliminary Survey Findings





2021-05-06 19:17:24 2826 0

When compared to Western regions, the Middle East North Africa (MENA) region has a unique artificial intelligence (AI) exploitation that is relative to its socio-economic situation and demographic structure. To get a comprehensive look at how MENA countries are implementing AI, the ​Access to Knowledge for Development​ (A2K4D) team created a survey targeting AI and data stakeholders in the region. The purpose of this survey was to explore AI adoption, production, benefits, challenges, capacities necessary and to understand how important data is in the AI and new technologies ecosystem, within the framework of its ​OD4D​ project. General Overview Our preliminary survey findings illustrate a broad context of the state of AI, showcasing where the region is as a whole in 2020. All respondents who took part in this survey had at least a basic understanding of AI. We managed to survey over 50 people across four countries: Egypt, Tunisia, Jordan, Palestine, and Lebanon. To get a better understanding of certain background characteristics of our audience, we asked for demographic information: 52.46% of the respondents belonged to the 18-24 age group. This is an indication that AI is most prevalent among the massive cohort of young people in the MENA region. It was also important to note respondents’ educational background; 48.2% indicated that they had a STEM (Science Technology Engineering Mathematics) educational background. The remaining 51.8% indicated non-STEM and specified businesses, tourism studies, journalism, social science, and economics. To determine stakeholders’ role in the ecosystem, we asked for respondents’ current occupation. Responses were as follows: 44.4% in academia (student, researcher or professor), 31.2% in the private sector (Business/SME/Startup founder or employee), 13.3 % in the public sector (Employee, policy maker), and 11.1% responded as independent (independent data scientist/ entrepreneur). This is an affirmation that in MENA, AI-initiatives are more led by non-government organizations. An AI-engaged region To evaluate the level of AI-engagement present and forthcoming in the region we asked our respondents if they are involved or plan to become involved in the field of AI (production, use, adoption research, or policy). While some stakeholders indicated that they are currently not involved in the field of AI (54.10%), more than three quarters (82.86%) of our survey respondents expect to be involved in the field of AI in the near future. This shows that the region is heading towards an AI-powered digital revolution. With machine vision and machine learning being the highest two in most popular forms of AI technology, there is a clear indication that efforts in technological advancements and innovation do exist. This is in line with the ‘Middle East & Africa Machine Vision Market 2019-2027’ report, which states that “The Middle East is increasingly moving from the oil & gas sector to other manufacturing & services sectors. With this, there is scope for other manufacturing and process industries investing in the region, thereby increasing the demand for machine vision in this region.” The MENA region is investing in modern automation solutions that are directed at upgrading new plants to compete in the global market. Though the MENA region is attempting to leapfrog into the digital transformation, it still lags behind from the rest of the world, as evidenced by our survey respondents. From our preliminary findings[1] we were able to point out suggested reasons as to why the MENA region is lagging behind the rest of the world: Both infrastructure and the legislative environment need improvement. There is a general lack of awareness on the potential of data and AI. Lack of appropriate educational modules to cope with new technologies. There is a huge gap between usage of AI in the public and private sectors. Lack of data openness and transparency; collecting big data remains an issue. Data being the major driver of AI adoption When we asked for respondent’s opinion on ‘how important data is to their work’ and ‘data is an important requisite for AI development’, 86% ranked data being the most important to their work and 92% strongly agreed that data is the most important requisite for AI development. We asked respondents to rank the difficulties they face that interfere with their data collection process, as an attempt to fathom the issues related to data quality and access in the region. Data access is cited as a constraint by our respondents: 23.3% believe that the legislative environment has negative effect on the data collection process; 20.6% believe that they cannot find what they are looking for; 19% cited data that is transparent or data quality, infrastructure was cited by also 19%; and data monopolies/ data ownership structures were cited by 18.1%. These ramifications are a result of the overriding issue of ‘open data’. Access to open data can unlock the potential of AI applications in the region, however, the region backslides because there is a general absence of both governments and regulatory organizations in this area. There is also a consensus that the absence of an open data legal framework will disrupt any efforts supporting and coordinating the use of AI in the region. Capacities Necessary Survey results show that investing in education and infrastructure are the main priorities needed to foster AI development. Furthermore, human resources and the legislative environment are also of great importance. AI 4 Development In the MENA, the private sector is ahead in terms of implementing AI initiatives and using new technologies[2], yet we wanted to further evaluate this statement. We asked for stakeholders’ opinions on AI adoption:Stakeholders believe that academia/research institutions and the public sector both have a significant role in the adoption of AI technologies. Regardless, AI being most prevalent in the private sector, the public sector still should act as a regulator, supporter, and policy facilitator for AI adoption, in addition to expanding and investing in AI research. While richer GCC states are exploiting the benefits of AI in order to position themselves as digital innovation hubs, AI efforts in the rest of the region are more bottom-up, often tackling developmental challenges. In Egypt, Tunisia, Palestine, Jordan, and Lebanon, the most pressing developmental concerns identified are: poverty, healthcare, unemployment, water scarcity and education. And 33.3% of respondents believe that AI could play a role in mitigating issues in healthcare, followed by education (22.2%) and poverty (16.7%). Reflective conclusion Our results suggest that stakeholders in the region are already showing enthusiasm with AI. The future ahead is all about moving in the right direction with the pace of digital transformation. Today AI implementation is slow because of the issues related to legislation, infrastructure, and open data. There is a dire need for an alliance between the private sector, academia, civil society and public sector. The private sector should coordinate with the public sector to keep up with the digital revolution, educational modules that encourage youth to cope with new technology, and civil society to address developmental needs. [1] Based on regional interviews (OD4D Paper 2: A Study of Data and Artificial Intelligence Narratives in MENA) [2] According to preliminary findings and insights by regional interviewees (OD4D Paper 2: A Study of Data and Artificial Intelligence Narratives in MENA)