Artificial intelligence, a potent technology with the potential to transform the future, presents a myriad of challenges and risks alongside its promises. Among these challenges, a critical issue looms large—the absence of adequate regulation and oversight concerning the data utilized in training AI models. Enter the Transparency Coalition, a novel nonprofit organization based in Seattle, dedicated to tackling this very issue.
The brainchild of industry veterans and tech stalwarts Rob Eleveld and Jai Jaisimha, the Transparency Coalition traces its roots back to a casual fireside conversation on Whidbey Island. Concerns regarding AI transparency and the unbridled use of training data spurred the duo to establish this nonprofit entity. Their core objective revolves around advocating for policy changes and fostering public awareness to foster a more ethical and accountable AI development landscape.
Both Eleveld and Jaisimha boast extensive backgrounds as technology and startup luminaries. Eleveld, a former U.S. Navy submarine officer, steered companies like Ekata and Whitepages, while Jaisimha, a University of Washington alumnus, honed his expertise at RealNetworks, Amazon, and Microsoft. Together, they bring a wealth of experience to the table, driving the mission of the Transparency Coalition forward.
Their current focus spans two pivotal areas:
-
Influencing Policy: Through active engagement with legislators in Washington and California, the Coalition seeks to shape state-level policies via advocacy and educational initiatives.
-
Educational Outreach: By enlightening stakeholders such as policymakers and business leaders, they aim to elevate awareness and understanding of AI-related issues in the broader public sphere.
The potential impact of mandating transparency around training data and model usage is profound. Such measures could refine the scope of AI applications, necessitating focused and consent-driven datasets to ensure compliance with copyright and privacy norms. This shift could lead to more predictable and accountable AI outputs, fostering a controlled and traceable ecosystem.
Envisioned legislation might encompass:
- Establishing standardized definitions for key AI terms.
- Requiring transparency regarding data used in model training.
- Implementing audit mechanisms to validate training data.
- Mandating opt-in consent for personal and copyrighted data usage.
On the financial front, Eleveld and his spouse kickstarted the Transparency Coalition with seed funding. As a 501©(4) nonprofit entity, the organization enjoys greater flexibility in advocacy efforts. Future funding pursuits involve seeking grants from foundations and entities keen on influencing policy in the AI domain.
Collaborations with AI research entities like the Responsible AI Systems and Experiences group at the University of Washington underscore the Coalition’s commitment to leveraging best practices and fostering dialogue between policymakers and AI experts.
In Eleveld’s words, demystifying AI models and dissecting them into comprehensible components is key to steering policy discussions in the right direction. By unraveling the intricacies of AI systems, stakeholders can pose pertinent questions and craft informed policy stances.
Audio editing and production by Curt Milton.
