From closed doors to open source

Chyung Eun-ju

Joel Cho

By Chyung Eun-ju and Joel Cho

Amid recent news surrounding the constantly evolving world of artificial intelligence, Meta has released an open-source AI software that the public can use to build their own AI chatbots.

Simply put, Meta's decision to make available its AI as open-source software means that it has borne at least part of the computer code behind its AI technology for free use and adaption by users.

With this development, Meta's competitors, like OpenAI, Microsoft and Google, have already voiced concerns and criticism of this move from Facebook's mother company. But are these concerns well based or do they stem from the leveling of the competitive field in the AI market?

Meta has taken a stance that providing its open-source software is a move that has more positive outcomes than negative ones, arguing that the secrecy kept by its competitors could actually hinder the general acceptance of AI technology by not only the public but governments as well. By keeping AI software closed to the public, the few giant players in this market could consequently possess complete control over this exciting new technology.

Here we will put aside the relevant and extensive discussion involving the need for government intervention and regulations for the sake of analyzing the impacts of Meta's decision to open-source AI technology on its competitors and the market as a whole.

Delving into the positive impacts of Meta's decision, it is undeniable that having open-source AI software from such an influential and important player in the tech world helps level the field for other smaller developers to dig their hands into the evolution of AI. This knowledge-sharing move results in a more collaborative and innovative market, since it allows for the growth of collective intelligence regarding the field and increases the potential acceleration of the development of AI technologies.

Open-sourcing software and its contribution to creating an innovative market is nothing novel. Take Mozilla Firefox for example. Mozilla made its web browser available as free and open-source software that disrupted the browser market and led to a more innovation-driven one as it allowed community-driven development, enhanced user experience for browsing the web by allowing customization and implementation of extensions that allowed users to meet their specific needs or preferences, extending what was at the time the traditional browser's capabilities. This innovative environment allowed developers to freely explore new ideas and implement experimental tools and features that refined and helped form the browser technology we have today.

By sharing the underlying code behind their AI software, Meta creates an environment that encourages developers, researchers and enthusiasts to contribute their expertise and ideas, and make improvements to AI technology. This collaborative environment creates opportunities for new applications and advancements in AI algorithms.

In other words, the availability of Meta's AI software allows the public to participate in the historic moment we are currently living in ― the evolution of AI ― driving innovation and expanding the boundaries of what is possible.

For instance, researchers at Stanford University developed an AI chatbot at a cost below $600. Nomic AI's CEO, Brandon Duderstadt, created a research product called GPT4ALL that empowers users to operate ChatGPT-like chatbots directly on their computer. Another researcher discovered a way to execute the software on a low-cost Raspberry Pi computer, showcasing the feasibility of developing chatbots utilizing large language models on low-cost hardware.

It goes without saying that Meta's decision to open-source AI technology is not purely altruistic as it benefits itself as well, establishing themselves as the thought leader and industry influencer in AI, since their decision goes against the secretive direction that has been the norm for its competitors.

However, it is also important to acknowledge that Meta's decision also raises concerns for its competitors. As mentioned before, the availability of open-source AI software levels the playing field and potentially increases competition within the market. So competitors may worry about losing their market share acquired through hefty investments in AI technology and furthermore, they may lose their competitive edge, as the necessary code behind AI technology becomes more widely accessible.

This sudden change in the competitive field will require that all players in the AI race adapt their strategies to differentiate themselves beyond the core AI technology in order to maintain their market position.

Apart from the impact on the competition and the market as a whole, once we dive into concerns revolving around social matters, Meta's decision to open-source its software could become a catalyst for the increase of the risks that have been in the centerfield for concerned experts.

With AI technology becoming more accessible to the public, AI systems can be used to spread biased information, discriminative and hateful views. Furthermore, at the stage, we find ourselves today regarding AI regulations, its scarcity allows the amplification of existing societal biases. Especially when it becomes about regulating tens of thousands of people that are strangers.

Chyung Eun-ju (ejchyung@snu.ac.kr) is studying for a master's degree in marketing at Seoul National University. Her research focuses on digital assets and the metaverse. Joel Cho (joelywcho@gmail.com) is a practicing lawyer specializing in IP and digital law.