Public Media Inclusion and AI

Tips for building trust and fostering belonging for Latinx audiences

The July 7 OIGO — which covered public media, generative artificial intelligence and opportunities for diverse engagement — prompted many conversations. I’ve been asked to say more. 💬

Advances in generative AI are creating new opportunities, and new dilemmas, for public media organizations hoping to better engage with and represent diverse audiences, including Latinos. 🤔 Now, public media is taking greater notice.

You may have seen and heard recent stories from NPR (and also) on generative AI and inclusion. 🔎 American Press Institute also tackled the subject of artificial intelligence from a publisher perspective. Beyond the ethical considerations I previously raised, what are public media’s practical guardrails?

In an era of rapid technological advancement, public media organizations are trying to keep up with information about generative artificial intelligence and its long-term effects on content creation and audience engagement. 🔊 While AI holds immense potential to expand services to diverse audiences, its integration presents unique moments when targeting Latine communities.

Are you considering AI as a way of addressing your diverse audience service needs? Here are some issues to keep in mind. 👇🏼

The hopeful: Generative AI technologies may offer a revolutionary way to create content, enabling public media to produce a wider array of stories, formats and languages. 💻 The power to customize content based on individual preferences holds the potential to attract and engage Hispanic audiences that have historically been underserved.

The real: As tools like ChatGPT gain widespread adoption, public media leaders must thoughtfully address critical questions. We know that generative AI models reflect imperfect human-generated training data. 😠 Thus, we risk perpetuating harmful stereotypes, prejudices and disinformation when we’re not vigilant in monitoring output and using training modalities.

So, if you’re considering generative AI for diverse content: Public media content and news staff will need to be active in addressing cultural sensitivity and language barriers. 📓 After all, the goal is to harness AI's potential to create authentic, relevant and meaningful content that resonates with Latine communities.

Here’s a quick checklist of areas to watch for building on belonging and trust with Latinx audiences. As well, this list may help you in maintaining public media’s broader commitments around diversity, equity and inclusion in our communities.

• Cultural Representation: AI-generated content must navigate the fine line between cultural appreciation and appropriation. 🧭 Ensuring that generative AI content represents the nuances of Latino/a culture and portrays it respectfully is a significant challenge.

• Language Barriers: Public media's commitment to serving Latinx communities hinges on accurate translation and language usage. AI-generated content must be able to convey the subtleties of various Hispanic languages to create a deeper feeling of engagement. 💭 You want the audience to feel seen and known.

• Diversity Within Communities: Latine audiences encompass a wide range of cultural backgrounds, each with its own traditions, customs and preferences. Content that your organization receives via generative AI must be adaptable enough to cater to this diversity while avoiding stereotypes. 👐🏾

• Relevance: I suspect most algorithms may miss the context of cultural references, idioms and historical events that are meaningful to Latina/o audiences. 📻 Content generated without proper context may appear insensitive or irrelevant.

The TLDR of this checklist is to avoid overreliance on technological solutions right now. Generative AI should complement, not replace, inclusive and intelligent content judgment. 🧐 When you directly involve Latina/o creators, experts and community advisors to guide appropriate use of generative AI, you build trust and reflect the diverse cultural needs of your area Latinx listeners and viewers.

Other methods for building your public media fluency:

• Diverse Teams: Public media organizations must ensure teams are culturally diverse and represent the communities we aim to serve. 🙌🏼 I know this is a tough one. I’ve covered inclusive hiring in a past OIGO, but happy to talk with you about it too.

• Training: You needn’t be an expert, but some base knowledge is great. 📖 As I’ve written before, I think the opening is figuring out how technology can benefit and strengthen us.

• Clarity About Human Oversight: Human review right now is critical. Content generated by AI is not as great as it will be in five to 10 years, and should undergo rigorous review to ensure it aligns with cultural authenticity, particularly. 💪🏿 But it’s okay to take that position as a diverse community service need.

• Collaboration: Partnering with Latine communities and other stations can provide valuable insights to inform our thinking on AI-generated content.

Another angle? Dick Tofel is among many people to offer strategic questions you should also consider. I believe there are diverse service angles to this too.

API suggests the best use of generative AI for now is outside of content, specifically around operational needs, such as sourcing job descriptions. 😓 This is especially helpful if your organization wants to attract diverse candidates, and hopes to ferret out barriers or biases in your hiring practices and JDs.

Intentemos esto… 👇🏾

I copied a job description and went to the large language model Claude. 📁 I then entered the following prompt:

"Please review the following job description for a [role] position at [public media organization]. Analyze the language used and identify any components that could present biases or inadvertently discourage diverse candidates from applying. Then, rewrite the job description using more inclusive language to attract a diverse pool of applicants. Focus on removing biased language, emphasizing required skills over credentials, highlighting our commitment to diversity, equity and inclusion, and framing the role as an exciting opportunity to make an impact.”

I pasted, and saw some interesting results. 👀 Give it a shot and see what you get. (By the way, I chose Claude because it does not have the character/word limits of GPT3, though you’re welcome to try with GPT4 access.)

Tofel forwards four questions you may want to consider. I’ll add to each some dimensions related to Latinx audience service. ⏰

1. Do we have the right talent?
   We’ll need technologists, creatives and community experts who deeply understand Latine perspectives to guide our AI understanding.

2. Do we have the right mindset?
   There’s openness in public media to experimentation and innovation, but we’ll need to find ways of balancing that with managing possible risks and harms to audiences that already face harms from media.

3. How fast is fundamental change coming?
   Generative AI is seeing rapid adoption, but could either supplant creation or be the next smart speaker. We’ll need to center Hispanic community trust.

4. What is the law—and what should it be?
   No clear legal framework exists yet to govern transparency, so we should lead by establishing ethical principles for serving diverse communities.

Public media's journey into the realm of AI presents difficulties, especially as we aspire to expand services to Latina/o audiences. In navigating these challenges, public media can redefine the way it connects with its diverse audience. There are spaces to foster a stronger sense of inclusion and representation in the digital age. 🟢

Cafecito: stories to discuss ☕

• More on the AI conversation and diversity: Researchers tested seven plagiarism detectors on non-native English essays. 🖥️ The verdict? 61 percent were falsely flagged as AI-written. The Markup explores more.

• Over 63 million Hispanics now live in the U.S. That's almost one in five Americans. 📊 Compare to 1980, when the percentage was just 7 percent. Now, Pew Research has released new fact sheets profiling how diverse our community has become.

  • There’s also a spreadsheet of detailed data.

• When disasters strike or new policies drop, some social accounts push lies in Spanish. They say climate change is a hoax, or that a research station controls the weather. 🤯 Most of this misinformation flies under the radar because moderators often miss Spanish content. NBC has details.

• For three years now, ¿Qué Hay de Nuevo, New Hampshire? has engaged the local Latinx community. NHPR has since expanded Spanish programming and hired its first bilingual reporter. 😀 Daniela Allee shares lessons learned with the Local Fix.

• WBUR this week hosted small-group listening sessions to hear from the area Latino/a community about its coverage and ways it could better serve listeners. 📻

El radar: try this 📡

• Is there an exodus locally? A commentary in the Aspen Daily News claims educated Hispanics are leaving for opportunities in more inclusive areas. 🧳 Are college-educated, skilled Latina/os migrating from your region? If so, where are they going and why?

• Investigate Latinx home ownership. 🏠 Rising mortgage rates are pricing many first-time homebuyers out of the market. As KQED’s Blanca Torres, reports, for Latinos and Blacks, already facing wealth gaps, reaching the dream of ownership is increasingly difficult. How might your public media organization cover how homeownership disparities exacerbate inequities in wealth-building for local minority communities?

• Take small steps toward source diversity. Alaska Public Media wants to grow its base of Latine news experts. 📈 In a cohort reportback, their goals may embolden you to try it too. “[P]rioritize talking about source diversity in reporter-editor conversations, brainstorm thematic approaches to story ideas, start a source database and commit reporting staff to having two to five coffees every month with nonwhite sources (with the time and freedom to build relationships.”

• Cover stewardship programs and Latine inclusion. KUNR reported on the new Lake Tahoe Stewardship Plan aims to make the area more welcoming for Latinos. Partners seek to protect the environment and serve local communities amid two million annual visitors. 🚶🏿‍♀️ How could your public media outlet localize this story?

• Link up around Hispanic food. Yelp just posted its list of Latinx businesses to watch, including restaurants. 🎒 This seems like a good opportunity to spotlight places near you. Yelp is renowned for its impact on customer traffic, and

The next OIGO is here Sept. 22. I’ll be sharing with you some thoughts at the conclusion of the Public Media Content Conference that week in Philadelphia. If you’re coming, join me at the sessions “DEI 2.0: Where Do We Go From Here?” on Wednesday at 3 p.m. and the Thursday morning “Connecting with Latino Audiences” plenary. Also, if you’d like to catch up, want to be interviewed for a future OIGO, or need some advice, reach out to me anytime.

🥤 You can buy me a coffee if you’d like to support the newsletter. There are also sponsorship opportunities available.