Earned Podcast: Ashley Waxman, VP of Brand Marketing at CreatorIQ hosts Jasmine Wong, Security Engineer, Creator, and Founder of Bytes Over Tea
Every brand and marketer is testing AI right now. Few are asking the most basic question first: what information am I handing over to get this tool to work for me. That question sits at the center of security engineering, and it is exactly where Jasmine Wong starts every conversation about AI. She is a rare combination of security practitioner and creator, with a community of over 115,000 followers across platforms and a newly launched in-person community called Bytes Over Tea.
In this episode, Jasmine and host Ashley Waxman, VP of Brand Marketing at CreatorIQ, dig into how her engineering background shapes the way she evaluates AI tools, why fear-based security messaging backfires, and what brands consistently get wrong when they use AI to work with creators. Along the way, she shares the habits she uses to protect her own data and makes the case for why the human connection is the one thing AI cannot replicate.
Watch the full episode
Evaluating a tool starts with asking where your data goes
Before she considers what an AI tool does, Jasmine asks where her data is going. Brands and marketers often move fast to adopt new tools without knowing who has access to the information they upload, where it gets stored, or whether it can be retrieved later. It is the first checkpoint she applies to any tool, especially when a brand feeds a confidential brief into an AI system without confirming how that data gets handled downstream.
There is also a trust layer that comes with being a creator. Her audience relies on her recommendations, so she has to stay conscientious about which tools she promotes and how she talks about them.
Fear-based security messaging discourages people from asking questions
Fear gets attention. Jasmine understands why so many headlines lean on it so heavily, but she has also found that fear makes people less likely to engage with a topic they already find intimidating. There is a second, quieter cost too: fear-based marketing can convince users that buying a tool eliminates all their risk, when no security posture ever actually reaches zero.
Her approach is to acknowledge the real risks directly, then pair every warning with a concrete takeaway and an action the reader can actually use.
AI still needs a human checking its work
AI, in Jasmine's view, is a tool that empowers people, provided the user can clearly articulate the input, the desired output, and what needs to be checked along the way. She has noticed that most people assume a single prompt will produce consistent results forever, when in practice tools require ongoing tuning, especially early in building a workflow.
That shift in responsibility is new. Buy traditional software, and a vendor fixes what breaks. Use AI directly, and the user becomes responsible for catching and correcting the output themselves.
Brands need to treat creator data with the same care as their own
When brands work with creators, Jasmine wants them to recognize what is actually at stake. Creators put their entire reputation on the line with every partnership, and that includes handing brands sensitive information like bank details for payment. She wants clear answers on how that information gets stored and handled, well before she sends anything over email.
She has also encountered brands that send AI-generated feedback on scripts she has written herself, feedback she can spot instantly and declines to use. Her audience, she points out, would spot it too.
Audiences can tell when a brand or creator has outsourced the relationship to a bot
Personalization is the tell. Jasmine has a clear signal she uses to judge whether a brand partnership is worth engaging with. When an email reads like automated spam, restarts the thread every time she replies, or lacks any specific detail about the campaign, she assumes it is a bot and does not respond. She has also seen brands generate a full slide deck and brief that directly contradicts the details in their own outreach email, forcing her to sort through outdated information before getting a straight answer.
Her guidance is direct: AI can make work more efficient. Real collaboration between creators and brands still depends on the human connection behind it.
The fix is training your AI in your own voice, then keeping a human in the loop
Using AI responsibly, for Jasmine, means spending real time training a model on how you communicate and what your expectations are, then building in checkpoints before anything gets sent. She built her own email-drafting workflow over months of training and still proofreads every message before clicking send. That same discipline, a human in the loop at the final step, is what she wants brands to apply before they push AI-generated content out to a creator or a customer.
She also uses AI extensively as her own operations team: automating post-metric tracking, packaging her negotiation history into a skill that recommends her rates, and capturing analytics screenshots for brand reports, all while keeping herself in the loop to review the final output.
Creator marketing works when brands give up some control
Brand awareness, not a single conversion, is what Jasmine tells marketing teams to build toward first. One video from one creator rarely drives a purchase on its own, especially for complex B2B tools, but it builds the recognition that gets a company onto a buyer's shortlist. She encouraged brands to give creators room to adapt messaging for their specific audience instead of expecting every creator to execute the same generic brief. A recent partnership stuck with her: a brand recognized her audience skewed toward cybersecurity and adjusted the brief accordingly, exactly the kind of collaboration she wants to see more of.
Security is everyone's department now, and Bytes Over Tea is built around that idea
Bytes Over Tea, Jasmine's new in-person community, brings AI users, builders, and security practitioners together for hands-on sessions covering everything from prompt writing to password hygiene. The idea behind it is simple: hearing "don't do this" once rarely changes behavior. Working through the habit in real time, alongside peers, does.
Her rapid-fire advice for marketers this week: keep work information out of personal AI tools, redact confidential details before uploading anything, and audit which apps have access to your accounts, removing anything you no longer use.
It is a fitting close for a conversation that keeps circling back to one idea: AI accelerates outcomes, but the relationships, the trust, and the judgment calls in between still come down to humans.
Listen to the full episode
And explore more conversations with the leaders shaping creator marketing on the Earned podcast hub.
Keep up with new episodes of Earned by following the podcast on Spotify, Apple Podcasts, and Google Podcasts, or if you prefer to watch these interviews, by subscribing to our YouTube channel.