Los Angeles CD 3, CA
Enhancing Language Access through
Community Based and AI Approaches
Los Angeles is one of the most linguistically diverse cities in the United States and faces significant challenges in providing equitable language access, especially for Indigenous language speakers. This diversity and lack of data surrounding indigenous languages creates communication challenges, hindering knowledge and access to essential news, events, resources, services and processes.
Due to staffing shortages and the complexity of managing diverse linguistic needs, CD3 struggles to deliver public service announcements, notices, and emergency information in all emergency response languages. This limitation disproportionately affects communities with limited English proficiency, hindering their access to vital services and resources.
Our project aims to support linguistic and cultural preservation by promoting equitable access to translation services across various fields and disciplines. This project builds on the Citywide Language Access Plan (LAP) to evaluate community-based methods and approaches for more accurate and culturally sensitive translations. Furthermore, the project investigates the role of AI-driven tools for meaningful language access improvements. By doing so, we seek to foster inclusivity and social participation, prevent the isolation of minority language communities, and support the survival and revitalization of these languages.
To gain a better understanding of the current landscape, we examined the policies, practices, and technologies related to language access. We began by reviewing Los Angeles' language access policies alongside approaches from other cities such as Boston, New York, and Seattle to identify best practices. We also evaluated relevant state and federal language access policies, including Title VI and Executive Order 13166. To explore technological solutions, we assessed AI-driven natural language processing (NLP) tools and community-based translation models. We analyzed existing translation tools like Google Translate, DeepL, Cohere Aya, and UNESCO’s NLLB as case studies to evaluate the current and potential roles of technology in translation services.
In the stakeholder engagement phase, we conducted interviews with staff from LA CD 3, representatives from community organizations such as CIFD, language experts from organizations like CIELO, researchers from academic institutions such as UCSC, and technology professionals from AWS to gain insights and perspectives into the significance of community collaboration and the application of AI language tools.
Language Access Approaches and Solutions: Policies, Partnerships, and Practices
This deliverable provides a roadmap for addressing language equity gaps and ensuring access to essential services in linguistically diverse communities. It includes a curated partnership directory of community-based organizations (CBOs), universities, and philanthropic entities to drive collaboration for a more inclusive and responsive language access framework. This collaborative approach facilitates a holistic, community-centered model for translation services, promoting mutual trust and respect for cultural nuances. Translation approaches for underrepresented languages, case studies of successful translation projects, and recommendations for contractor engagement are outlined to further support practical implementation. Additionally, relevant state and federal policies are mapped to establish a legal and operational foundation.
AI Language Access Implementation Toolkit
This toolkit evaluates the role of AI and technological tools to equip CD3 staff with the resources to make informed decisions about adopting translation technologies. It includes frameworks for assessing AI translation tools, criteria for vendor evaluation, and strategies for fostering community engagement. Practical frameworks were developed to guide the evaluation and responsible implementation of AI language technologies, emphasizing cultural sensitivity and community-centered approaches. Recommendations were provided to support evidence-based decision-making in AI adoption and to ensure alignment with community needs. The deliverable also facilitates effective collaboration between technical teams and community stakeholders, promoting inclusive and ethical integration of advanced language translation technologies.
By strategically addressing language equity gaps, the city can increase its capacity to protect and promote linguistic diversity, particularly as Los Angeles grapples with resource limitations while striving to serve its increasingly diverse and multicultural population. Providing language access helps foster social inclusion, ensuring that Indigenous groups and non-English speakers can advocate for their needs and fully participate in civic, social, and economic life. It enables Indigenous communities to maintain their cultural identity while also integrating into broader society, ensuring their voices are included in shaping policies that affect them. This project supports the city’s efforts to build a more inclusive, just, and engaged community.
By adopting these recommendations, CD3 can improve linguistic inclusivity, enhance access to critical services, and preserve the cultural heritage of Indigenous communities. Future work includes:
1) Testing AI tools with community feedback to refine translation quality
2) Expanding partnerships with regional and national organizations
3) Monitoring the impact of implemented solutions through standardized metrics
Wendy Zeng (Project Lead)
UC Berkeley
Smita Samanta (Project Lead)
New York University
Emma Christie
UCLA
Asli Kocak
Tufts University
Antara Chugh
UCLA
Karina Castro
San Jose State University
Elizabeth Fonseca
UC Santa Cruz
Shreya Taluja
University of Pennsylvania
Matheus Dias
UC Berkeley
Yasmin Jafari
UCLA
Josh Crowley
Pomona College