Prime Minister has unveiled Vietnam’s ambitious artificial intelligence masterplan up until 2030: Vietnam’s 2030 targets will take it into ASEAN’s top four:
- following the Oxford AI country Readiness index, Vietnam is currently ranked in 76th position in the World, 6th in ASEAN with 42.944 pts behind Singapore (6th, 78.704), Malaysia (28th, 62.434), Thailand (60th, 48.156), Indonesia (62nd, 47, 528), Philippines (74th, 42.944)
- and the world’s top 50 countries in terms of AI research, development, and application.
The government wants Vietnam to establish 10 reputable AI brands or services by 2030 and plans to build 3 national big data storage centers to give businesses access to high-performance computing.
Vietnam is still lagging behind many countries, are these objectives realistic?
If we are to assess the situation today, to know the strengths and weaknesses and to reinforce or improve them, we need to have metrics, calculated periodically, allowing us to compare with other countries and to monitor progress, and to know the remaining path to achieve these goals.
Existing AI indexes
There are a few indexes on Artificial Intelligence, based mainly on the results of surveys and published annually.
Vietnam is present, and relatively poorly ranked in some of them:
- Global Government AI
- Readiness Index, 2020
- Global Cities AI Readiness Index, 2020
- Cisco Digital Readiness Indexes, 2020
- Digital Adoption Index (DAI), latest 2016
- IBM/Watson Global Adoption Index, 2021
- Global AI Adoption Index, 2021 (IBM/Watson)
- The Inclusive Internet Index, 2021
- Digital Economy and Society Index (DESI), 2020
- Digital Society Index (DSI), 2020
- AI Watch Index
- AI Hiring Index
- Digital Intelligence Index
These indexes shed light on certain particular points such as the role/strategy of government, the adoption of IA, or the labor market, but there is not yet a more global index combining these factors.
It would be interesting to have a family of indexes quantifying the motivation of the government and its measures put in place, the adoption by the general public, and enterprises, and the impacts and the results on the labor market, and the benefits on cost, revenue, productivity and efficiency.
COVID-19 is also a boosting factor of AI, especially for Manufacturing, Commerce, Medical, Education, Finance.
Criteria could be classified into pillars (non-exhaustive list):
- Government/Regulation: Vision, Governance & Ethics, Digital Capacity, Adaptability, AI Risks and Penalties (see EU regulations)
- Data and Infrastructure: Infrastructure(Cost, Performance, …), Data availability (History, Volume, …), Data Representativeness
- Demographic/Economic: Population/Youth, Internet/Mobile Users, Labor, Diversity, …
- Technology Sector:
- Size (investments, projects value, market capitalisation listed companies),
- Innovation Capacity,
- Human Capital,
- Job Marketplace (Offer/Demande, Salary average vs other IT),
- Education/Training & Research,
- Competition between domestic vs foreign AI companies located in Vietnam (attractive salary, working conditions, …)
- Business Profitability: Listed companies prices returns/performance like IFRC Global AI indexes …
- Adoption
- Survey: Adoption/Transformation/Impacts in industry, enterprise (see study by McKinsey), population/everyday life, COVID-19, …
- Implementations in different area, Tools/Apps/Libraries
Occurence keywords in social networks, online newspapers, …. - Benefits, Impacts and Results of AI Statistics
Vietnam AI Development Indexes (VAIDI)
The quality of the indexes obviously depends on the input data.
These data and statistics can be obtained from free, public, or paid sources. Like the other existing indexes, the results of surveys will also be essential.
The construction of the index (and sub-indexes) must be transparent, replicable for any country, region, city or sector. It will be a weighted average (equal, different, or optimized after research) of the ratings of the various criteria. These criteria are quantified, rebased between 0 and 100.
They will make it possible to evaluate the strong points and the weak points, to reinforce strong points, and to improve weak points. These input/output data and indicators can then be integrated into explanatory/forecasting models for international scientific publications.
