White House Enters Debate on “Open” Versus “Closed” AI Systems
According to the AP, the Biden administration has stepped into the ongoing debate surrounding the openness of artificial intelligence systems, sparking discussions among tech companies, researchers, and data professionals.
The White House announced its intent to solicit public opinions on the implications of making critical AI system components open-source, allowing for widespread access and modifications. This initiative forms part of a broader executive order signed by President Joe Biden in October, aimed at steering the rapid advancements in AI technology.
The discourse on AI openness is polarized. Tech giants like Facebook’s Meta Platforms and IBM champion the cause for transparency, while others caution against the potential hazards of making AI model components readily accessible.
President Biden’s executive order has highlighted the need for a deeper investigation into “dual-use foundation models with widely available weights,” referring to the numerical values that dictate AI model performance. The openness of these models could foster innovation but also pose security threats, including the dismantling of built-in safety measures.
The Commerce Department’s National Telecommunications and Information Administration, or NTIA, is facilitating 30 days for public commentary, aiming to gather diverse insights for a comprehensive report.
Alan Davidson, NTIA’s administrator, emphasizes the complexity of the issue, suggesting that the debate is not merely binary but encompasses various degrees of openness, indicating a potential pathway to balance innovation with security.
In alignment with the administration’s open call, Meta has expressed its willingness to share insights from its decade-long experience in developing AI technologies openly. Conversely, Google, which generally advocates for a more guarded approach, recently introduced Gemma, a suite of open models derived from its Gemini technology.
This move suggests a shift towards transparency, albeit with caution, as Google underscores the importance of a nuanced perspective on the open versus closed debate, urging for a collaborative effort to address potential risks and benefits.
Critics argue that merely opening AI components does not guarantee accessibility, as the effective use of such models still demands significant resources, predominantly held by a few major corporations.
David Gray Widder, a researcher from Cornell University, points out that the motivations behind advocating for open or closed models are multifaceted, with each stance potentially serving the interests of those who adopt it.
As the Biden administration ventures into this contested terrain, the call for public input signifies a pivotal moment in shaping the future of AI development and regulation. The outcome of this inquiry could redefine the balance between innovation and security, impacting the trajectory of AI technologies and their governance.
Originally posted on OpenDataScience.com
Read more data science articles on OpenDataScience.com, including tutorials and guides from beginner to advanced levels! Subscribe to our weekly newsletter here and receive the latest news every Thursday. You can also get data science training on-demand wherever you are with our Ai+ Training platform. Interested in attending an ODSC event? Learn more about our upcoming events here.
