Gareth Watkins - PrideNZ and Artificial Intelligence

The audio recording features an in-depth discussion between Gareth Watkins, founder and director of PrideNZ, and Roger Smith. Conducted on December 15, 2024, in Wellington, Aotearoa New Zealand, the interview delves into the origins, objectives, and evolution of PrideNZ, with particular emphasis on the integration of artificial intelligence (AI) into its operations. The recording highlights the pioneering steps PrideNZ has taken to preserve and promote rainbow history and discusses the challenges and opportunities that AI presents for archival work.

Founded in 2009, PrideNZ has steadfastly maintained its commitment to three core goals: serving as a first-hand community resource for rainbow stories, promoting inclusivity and well-being, and remaining a not-for-profit initiative. Over its 15-year journey, the organization has built an extensive archive of over 1,000 audio recordings, encompassing interviews, community events, and commemorations such as the annual AIDS Candlelight Memorial. Watkins emphasizes the importance of accessibility, ensuring that these recordings remain public and serve as an educational tool for future generations.

PrideNZ’s collaboration with major institutions like the Library of Congress and New Zealand’s Alexander Turnbull Library underscores its role as a custodian of cultural heritage. In 2021, the Library of Congress expressed interest in capturing the PrideNZ collection for posterity, marking a significant milestone. This recognition spurred further partnerships, particularly with the Alexander Turnbull Library, leading to a systematic monthly deposit of audio recordings into their archives. Watkins views this collaboration as crucial for ensuring that rainbow narratives are represented in state-run institutions, fostering greater visibility and preservation.

AI has become a cornerstone of PrideNZ’s efforts to make its archives more accessible. In 2020, the organization began using transcription tools like Otter.ai to automate the labor-intensive process of converting audio into text. This shift not only reduced costs but also enhanced discoverability by enabling search engines to index the content. Watkins describes the progress in AI’s transcription accuracy, noting significant improvements in recent years, though challenges remain, particularly with Te Reo Māori and accents from Aotearoa.

The integration of AI extends beyond transcription. PrideNZ uses tools like OpenAI’s API to generate summaries of audio recordings, creating concise, searchable descriptions of lengthy materials. This automation allows the organization to process its extensive catalog efficiently, making previously inaccessible content available to a broader audience. Transparency is a priority; all AI-generated content on the PrideNZ website is clearly labeled, and users are encouraged to verify summaries against the original recordings.

A notable ethical consideration is the use of pronouns in AI-generated content. Given that PrideNZ does not track the pronouns of contributors, summaries default to gender-neutral language. Watkins emphasizes respect and adaptability, allowing individuals to request updates to ensure accuracy.

Watkins highlights the transformative potential of AI for small organizations. Beyond transcription and summarization, tools like ChatGPT offer opportunities to create interactive experiences, such as chatbots capable of discussing rainbow history. By enhancing accessibility and discoverability, AI has elevated PrideNZ’s visibility, enabling users to engage with its content through multiple channels, including the National Library’s platform.

Looking ahead, Watkins identifies data quality as a key priority. Accurate metadata—such as dates, locations, and participant details—is essential for maximizing the effectiveness of AI tools. PrideNZ is also exploring new applications for AI, such as summarizing historical newspaper articles.

Watkins’ dual role as Collections Data Manager at Te Papa further informs this work, offering insights into how AI can be leveraged across different types of collections. He advises other organizations to experiment with AI, emphasizing the importance of starting small, using public data, and gradually building technical expertise.

This summary is created using Generative AI. Although it is based on the recording's transcription, it may contain errors or omissions. Click here to learn more about how this summary was created.

Gareth Watkins

2000s, 2010s, 2020s, ai prompt, alexander turnbull library, application programming interface (api), archives, artificial intelligence, chatgpt.com, collection management system, descript.com, gareth watkins, generative artificial intelligence (gen ai), google.com, ia, interviewing, library of congress, museum of new zealand te papa tongarewa, national library of new zealand, notebooklm.google, openai, otter.ai, pridenz.com, pronouns, python, roger smith, te reo māori

accents, accessibility, advice, advocacy, aids, alexander turnbull, audience, authenticity, balance, bottom, change, christmas, code, community, conversation, data, discovery, email, energy, events, experiment, fear, fish, glam, health, heritage, history, job, journey, knowledge, language, library, literacy, memorial, metadata, newspapers, opportunity, other, otter, people, podcast, press release, rainbow, recognition, relationships, resilience, resource, running, stuff, time, trust, website, work