Image credit: The Edge Malaysia
The year 2023 has been a thrilling one for artificial intelligence (AI). Nevertheless, effectively managing the tension between the rapid progress of AI and the imperative for cautious approaches to minimise potential harm remains a central challenge.
In 2021, the European Union initiated the development of regulations for AI, signalling the first measures to tackle the challenges posed by this technology. Despite an early advantage, the implementation of these regulations is not expected until next year.
Following suit, Malaysia is also establishing a framework for AI governance and a code of ethics. This move is driven by the growing interest in AI businesses expanding their presence and entering new markets.
Malaysia is currently addressing this through the National Artificial Intelligence Roadmap 2021-2025, launched by the Ministry of Science, Technology and Innovation to enhance the country’s AI capabilities.
The road map incorporates the Seven Principles of Responsible AI, offering guidelines for the development of trustworthy and responsible AI to safeguard individual rights and privacy. This is crucial for establishing trust in AI systems where it hinges on the industry’s ability to operate reliably, safely and consistently, especially in critical fields like transportation, healthcare and financial services, where consequential decisions are made that affect both lives and livelihoods.
There is already a growing hunger for AI across a wide range of industries, developers and businesses. As enabling technology platforms, cloud providers should do what they can to help customers learn from and use the latest technology and tools. AI and, in particular, generative AI are possibly the hottest ticket in town, and we have seen phenomenal interest in our own large language model (LLM).
But we have to find a way to strike a balance between enablement, growth and safety. They are all connected and not exclusive, and open source has to be considered a valid proposition for what will be an ongoing debate about methodologies.
Alibaba Cloud has long been an advocate of open source and supporting transparency when applying AI. In 2022, we launched our own cloud-based open-source AI model community called ModelScope. The platform has provided 2.8 million developers, accounting for 35% of developers in China, with access to over 2,300 AI models within a short period of 12 months, making it the largest AI model community in China.
By December last year, Alibaba Cloud had already open-sourced its LLMs with parameters ranging from 1.8 billion, 7 billion, 14 billion and 72 billion, as well as its multi-modal LLMs with audio and visual understanding capabilities.
A principled approach
Without open source, we wouldn’t have had the ground-breaking development of generative AI but, of course, this is a fast-developing space where commercial propositions will grow and proprietary systems will emerge.
If IT history has taught us anything, it is that innovators and entrepreneurs will want to land grab ideas and software markets and niches. We are in the business of enabling that invention and development but believe that those ideas and products will be stronger and grow quicker through an open-source community.
Public sector and industry organisations will have to continue to collaborate on regulatory guidelines but, in the meantime, the industry continues to move forward at an incredible pace. In recognition of this, we have our own set of guiding principles for our AI development to make technology “available, reliable, credible and controllable”. We invested in technologies such as privacy-preserving computation and explainable AI to increase our algorithm’s transparency and fairness, protect our users’ privacy and enhance data security.
Earlier last year, we made an announcement that we aim to be the most open in the era of AI to make it easier and more affordable for everyone to develop and use AI. Generative AI is already being used across a range of organisations, from big luxury brands reshaping online retail experiences to digital intelligence and AI enhancing supply chain agility and resilience. We are also working with many industry pioneers and start-ups to establish next-generation experiences in gaming, music and retail.
The point is that ecosystems can drive innovation by using AI to enable businesses to think beyond their traditional boundaries. Open source prevents lock-in with proprietary systems and ensures a lower barrier to entry for innovators and start-ups. It encourages experimentation and collaboration at lower costs and within robust communities, opening up opportunities with less risk.
At a time when business talk is constantly touching on flexibility and resilience, open source can provide a willing platform to scale at cost. It almost certainly has to be part of the mix, a serious factor within a bigger conversation about the future of AI.
The challenge for all countries and companies is how to engage, enable and manage AI now. There is already significant momentum and the tech industry as a whole has to play its part in taking responsibility for being the guardians of AI for today and tomorrow.
Source: https://theedgemalaysia.com/node/705638