- The Upgrade
- Posts
- Rolli Founders on Combatting Deepfakes
Rolli Founders on Combatting Deepfakes
Plus, I polled media experts on AI, top headlines, and more!
Welcome to The Upgrade
Welcome to my weekly newsletter, which focuses on the intersection of AI, media, and storytelling. A special welcome to my new readers from Reddit, The Wall Street Journal, Columbia Business School, and many other top organizations — you’re in good company!
In today’s issue:
The Week’s Top AI Stories 📰
My AI Poll 📲: Insights from This Week in Digital Media
🎙️The Big Interview: Rolli Founders Combat Deepfakes
Special Course Offer: 15% Off the AI Boost⚡️
The Week’s Top AI Stories
Gen AI Tools
Everything announced at OpenAI’s first developer event — Tech Crunch
OpenAI Lets Mom-and-Pop Shops Customize ChatGPT — The New York Times
Elon Musk Shows Off New AI Chatbot by Requesting a Recipe for Cocaine — PCMag
Can an AI Device Replace the Smartphone? — The Wall Street Journal
AI Arrives on YouTube With Comment Summaries and Video Suggestions — Decrypt
Regulation & Policy
Big Tech wants AI regulation. The rest of Silicon Valley is skeptical. — The Washington Post
AI safety: How close is global regulation of artificial intelligence really? — BBC
Ethics & Safety
AI pioneer Fei-Fei Li: Give scientists more access to advanced AI models — Axios
Meta to Require Political Advertisers to Disclose Use of A.I. — The New York Times
We’re not ready for a major shift in visual journalism — Poynter (opinion)
Legal & Copyright
The New York Times got its content removed from a huge AI training dataset — Business Insider
AI will complement Hollywood, says CEO of Google-backed video AI startup — Yahoo Finance
AI is about to face many more legal risks. Here’s how businesses can prepare — Fortune (opinion)
In the Workplace
Poll Results from This Week in Digital Media
This past week, I had the pleasure of guest hosting This Week in Digital Media, an SMS-based community chat powered by Subtext. In addition to sharing my story and (shamelessly) plugging this newsletter, I sent numerous polls throughout the week to the list of roughly 700 media experts and enthusiasts. My goal? To learn their current sentiments regarding AI. Using Subtext’s powerful scheduling feature, I planned out the messages in advance, sat back, and watched the responses come in.
I first asked, are you more concerned, excited, or equally so about AI?

Here’s how the This Week in Media Poll compares to a nationally representative August 2023 poll from The Pew Research Center:

Then, I asked the SMS group what their biggest concerns were re: AI.

Unsurprisingly, there were a LOT of “other” reasons mentioned, including, in order, disinformation, ethics, quality, and global security.
The Big Interview: Rolli Founders on AI
Nick Toso and Catalina Villegas are the co-founders of Rolli, a curated platform to connect journalists seeking sources with experts seeking coverage. Rolli’s AI-enabled expert discovery process is designed to speed up newsgathering dramatically.
Peter: Tell me how Rolli serves journalists and experts and what that user experience looks like.
Nick: The easiest way to describe it is how our users describe it: LinkedIn for experts. You imagine a LinkedIn platform. It's very similar to that. Experts apply to join the platform, are vetted, and then upload information to their profiles, including their area of expertise, the events they attend, and the articles they publish or are featured in. Journalists then search organically to find the experts based on whatever they're working on that day.
Peter: You mentioned it's similar to LinkedIn in some ways. Why couldn't journalists try to reach out to sources there instead? There are tons of platforms — Muckrack and HARO — which all are trying to address this question and this problem. How is this any different?
Catalina: Essentially, what a lot of these PR companies do is they take our information, and they sell it. And they don't necessarily realize that that's not how journalists work. We aren't sitting by waiting for our inbox to get flooded with pitches, which is how the industry operates right now. The majority of these companies use email to send coverage requests to journalists. But in reality, the way journalists work is that I get up every single morning, I have a meeting with my producers, we talk about what's going on in the world, and we come up with an idea for what we're going to cover that day.
That's essentially what a journalist's life is like. According to the Society of Professional Journalists, 50% of our working hours are spent looking for experts. It's a monumental task, and it's really hard to do efficiently.
Peter: What’s broken about the current newsgathering process?
Catalina: When you have these PR companies pitching us, it's all coming in at the wrong time, it's not vetted, and it's often not what I'm working on. Rolli is unique because it works in real newsroom workflows. Many of our team members are still working in newsrooms, so we understand the problem deeply.
Peter: I'm curious about how the platform matches journalists and experts. How does your AI-powered algorithm work?
Nick: For journalists, what we're doing with AI is helping them find the absolute best interviewee possible. With AI, we're able to go past a simple keyword-based search for an expert. We have a logic-based search, which picks up the nuances of an expert's expertise. For example, if you were to go into another platform that has a keyword-based search and search special counsel, it would give you experts for the Special Olympics, the Council on Foreign Relations, or things like that. It wouldn't give you a lawyer. But with our AI, it understands that it's a lawyer you’re searching for.
Another example is you might have an expert, and somewhere buried deep in their bio is one line about how they helped an agricultural project in Kenya. It's one simple line. It wasn't in their title. It wasn't in their keywords. It was just in their bio. Our AI is able to pick up on those cues so that if a journalist goes into our platform looking for food supply chain expertise in Africa, this expert will show up because the AI recognizes that this person has experience working on food supply issues in Africa. And it's transformational, frankly. It gives you a much deeper understanding of an expert's expertise. It gets you who you want to talk with instantly.
Peter: How does the vetting process work for experts to join your network, and why is it so important?
Catalina: We have a seven-step vetting process that we have put together according to the best practices outlined by institutes like the Poynter Institute, the International Center for Journalists, and the Associated Press. But the most important way that we do it, of course, is from the get-go. Who are we reaching out to? Who are we inviting to join the platform? We take that extremely seriously.
Nick: Vetting is at the core of what we do here. We do the job that a journalist just doesn't always have the time to do anymore. Our work is only becoming more important in the age of AI deepfakes. For example, 44.7 million fake accounts were removed from LinkedIn during the second half of 2022. That's almost 50 million fake accounts that they detected during registration. And they’re about to get a lot harder to detect. We've been seeing bots on Twitter for a long time. Now, we're seeing all these new pieces of technology that can generate seemingly real people via imagery, audio, and video. And you could potentially even interview them without knowing they were fakes! That’s where Rolli comes in.
🎓 My live online class starts next week! 💻
AI Boost for Professional Communicators and Marketers covers the essentials of Generative AI for media and marketing professionals with novice and beginner-level experience with AI tools. The live 90-minute sessions will take place on Tuesdays, Nov. 14th, 28th, and December 5th, at 7pm ET / 4pm PT. Level up before 2024!⚡️
Don’t be shy—hit reply if you have thoughts or feedback. I’d love to connect with you!
Until next week,
Psst… Did someone forward this to you? Subscribe here!
Reply