My Leap into AI
At the end of 2023, I was laid off by The Verge and Vox Media just two months into my maternity leave. I was the only person impacted at The Verge and the only mom in leadership at the time.
While I think I’ll always be a little triggered by the experience - feeling kicked when I was down and vulnerable, I’ve not looked back. As Laura Brown says in her forthcoming book, All the Cool Girls Get Fired. And I couldn’t agree more.
If you read nothing else beyond this: move on and find new space to bloom.
Just 3 weeks later, I traversed the step and repeat at the National Daytime Emmys with my former boss, bleary-eyed from a lack of newborn sleep and postpartum hormones. But even then, I knew my aperture was widening. I knew that above all else, I was resilient, adaptable, and curious.
The show I executive produced lost at the Emmys that day, but that night, on the rooftop of my hotel, I wrote what has become a manifesto for myself: I am just getting started.
Before I left Vox Media, my teams had begun to incorporate AI tools into our production workflows, but I myself never had the time to meaningfully play with them. Drowned in meetings and oversight, I had become removed from any tangible creation and experimentation.
Flash forward to now: I’ve spent the last 18 months immersing myself in all things AI across text and search, audio, and video. I’ve solo-produced an AI-generated podcast, cloned my own voice, dabbled in agent creation, leveraged deep research and Notebook LM for all sorts of personal and professional purposes, painstakingly refined my own multimedia content prompts (understanding that this is my real asset), and recreated childhood memories using a slew of genAI image and video tools (hello, Freepik and Luma AI!). I’ve also jumped in on the ‘vibe-coding’ trend in Cursor and am hooked by what I can create myself. And though it's liberating, the practice deepens my respect for experts in these fields.
Along the way, I’ve been thinking deeply about the role of AI in journalism and nonfiction programming, and coaching myself to be comfortable with, and even thrilled by, the uncertainty and the unknown. More on this later!
I’ve also put myself out there, attending any and all AI events, conferences, and meetups in San Francisco and LA. I’ve advised the hugely ambitious AI Collective, participated in AI video hackathons with Machine Cinema, attended AI film competitions with Project Odyssey and Civitai, and had countless meetings with people just like me - curious and trying to figure out their place in this new era.
More than anything, I’ve become a student again, learning the marketplace for publisher content with companies like ScalePost so I can use my experience to thoughtfully address publisher challenges and help create a more robust product on the technology side. It is thrilling to be building a new model for news and information consumption as we are at Perplexity, and I am learning so much working again with product managers and engineers - it reminds me of my earlier days at Beme. I’m even learning about women-led venture capital thanks to Capital F and am inspired by the driven women putting something new into the world.
Here’s a few snippets of what I’ve learned in the last 18+ months:
GenAI tools are stronger and will have more adoption and longevity when there are creative and editorial voices helping to build and shape them. We need each other for this to work.
AI will never replace on-the-ground, human-led reporting and storytelling - an area where we desperately need more. (Though there are certainly exciting ways it can amplify it.) Effective deployment of AI should create more space for this vital work.
If we understand journalism as the vital dissemination of information for a functioning democracy, then for it to be truly effective, it must reach as many people as possible. This REQUIRES capturing attention through compelling execution and leveraging the most advanced tools of today. In this way, AI adoption and integration isn’t a nice-to-have but a mandate, requiring continual reinvention.
Integration into production workflows is the obvious first step. DeepResearch efficiently onboards you to subjects for faster deep dives. Upload preliminary materials to NotebookLM and take a walk to digest everything! Use it for interview prep - cross-check previous interviews and related multimedia content easily. Data analysis, research outlines, call sheets, production timelines, budgets - AI can create the foundation, with your close oversight, and it'll become even more seamless when agents with memory are incorporated. That's not even mentioning the exciting pre-viz, production and post tools available now. I'm particularly excited about AI editing tools like Eddie AI, that accelerate string outs and offer fresh options and missed material when you hit those inevitable narrative roadblocks.
With the above said, I'm seeing extensive discussion about AI ethics in reporting workflows, but minimal conversation around creative and storytelling applications. I firmly believe multimedia journalism desperately needs reinvention and deserves greater freedom to experiment with creative storytelling and AI video tools (with clear labeling)- the same latitude traditional documentaries enjoy (thanks Fred Grinstein for this astute observation). Here, I think of Joe Posner's early experimentation with AI at Semafor to imagine life in the early days of the Ukraine war and Bob de Jong’s archival pairings with AI recreations of historic Manhattan.
Start with a tool that is outside your skillset or specific expertise. If you’re a producer, try creating your own b-roll with Runway, Luma, or Kling to pair with your script. If you’re a shooter, try writing your own script with Claude and creating a voiceover with ElevenLabs (or clone your own voice!). The list goes on across traditional domains and production stages. In this way, you become your own creator - a foundational step in making AI tools work for you (and a tenet I will always thank The Verge for championing). Don't overthink it.
All journalistic practices and keen attention to detail still apply to any of the above! I am not minimizing the need for rigor in fact checking, spotting and addressing hallucinations, or the deleterious impact of deep fakes or misrepresentations. But I firmly believe that we shouldn't write the rules until we have tangible examples to point to, and in that exploration, we may spot creative lifelines for multimedia journalism’s next chapter.
There’s much more to share. In the meantime, I’m excited to be at AI on the Lot this week. If you’re there and want to talk AI + nonfiction, reach out.