The Malaysian government is in the midst of developing its own Large Language Model (LLM) to support the local artificial intelligence (AI) sector. This was revealed by deputy science, technology, and innovation minister Mohammad Yusof Apdal in parliament today.
According to the deputy minister, the ministry is working together with the health ministry, digital ministry, higher education ministry, plantations and commodities ministry and the prime minister’s department for this endeavour. The development will adhere to the government’s AI Governance and Ethics guideline to ensure that the LLM does not “ignore local values and culture”.
Yusaf stated that Malaysia’s own Large Language Model would reduce the reliance on external technologies. He also noted the high costs of the development process which can be reduced by using cloud service providers. This brings up data security concerns as data stored abroad will have a risk of confidential data being leaked, but he claims that through MIMOS, his ministry is exploring the use of computer infrastructures with a competitive cost.
Moreover, he added that the country is taking steps in developing an LLM in the Malay language through collaboration with local universities and companies. One of the project’s key partners is Mesolitica, which has already developed the Malaysia Large Language Model (MaLLaM) that was trained using local datasets. The deputy minister also said that the Dewan Bahasa dan Pustaka (DBP) will also be involved in the future.
(Source: Parlimen)
