Lucky Gunasekara is CEO and co-founder of Miso, an AI media lab which partners with more than 50 publishers globally to explore how the shift toward “liquid media” — adaptive, modular, and machine-readable content — is transforming the engagement, distribution, and revenue models of news and media. His current focus is on how autonomous agents, open-source models, and inter-operable web standards could enable new forms of media products, sustainable economics, and institutional resilience in an increasingly agentic media ecosystem.
Founded in 2017, Miso helps media and research businesses (including Flashes & Flames) turn their archives and specialist content into conversational answers, research tools and personalized reader experiences. Its flagship Answers product generates source-cited responses based exclusively on a publisher’s own material, rather than the wider web.
Gunasekara had graduated from Cornell University in neurobiology and behaviour, completed a Fulbright Fellowship at Kwansei Gakuin University, in Japan, and subsequently attended Stanford University School of Medicine. He has managed search and natural-language-processing research within a DARPA-funded laboratory and also worked at Cornell Tech’s Small Data Lab with Miso co-founder and CTO Andy Hsieh, exploring privacy-first personalization, contextual advertising and immersive search technologies.
At Miso, he combines product leadership with advocacy for a publisher-led approach to generative AI. He argues that AI systems should respect editorial standards and provide appropriate credit, consent and compensation to experts, publishers and creators. Miso powers billions of personalized reader experiences and millions of AI-generated answers each month for media companies worldwide. The co-founder is also involved in developing open, community-based standards intended to give publishers greater control, inter-operability and sustainable economics in an AI-dominated internet.
Lucky Gunasekara is among the 32 CEOs and industry leaders speaking at The Definitive AI Forum for Media, Information & Events, on 18 November in London. We asked him to predict the greatest AI challenges for these industries over the next 2-3 years. Here’s his reply…
Miso’s Lucky Gunasekara: up for the fight with ‘Godzilla‘
I am very concerned that Google will unravel further so anyone dependent on search referral traffic at an existential level is in for a very bad further shock. And I am alarmed by how many publishers are feeling these effects already, laying off staff, and resorting to AI automation to make it work – or just accepting smaller audiences and revenue.
That’s not a winning strategy. So, I expect a good deal of consumer media to be in for a lot of tumult in the next 2-3 years. I do not think programmatic licensing of publisher content will work. The AI firms don’t seem remotely interested in it. And the revenue-share models don’t make any sense. You’re talking, at best, a $15 CPM per grounding Answer with a 50-50 revenue share. So already Google, Microsoft, and Cloudflare take half of that pie. And then you’re left dividing $7.5 CPM amongst 10-20 of your publisher friends.
It might work for consumer media but it cannot math out for B2B subscription media. So, I think what we’ll see is licensing deals being chased by consumer media, while savvy B2B media groups will adopt MCP as a new subscription model for professional agents and sell that capability into their enterprise corporate accounts.
Why? There’s no revenue share: they retain control of their business and the data – and it deepens corporate sales as a revenue base. But I worry about the mania we have on programmatic licensing as a concept, and every time this topic comes up, we lose a year or two of momentum out of being distracted. It’s a mirage.
I think we’re also underestimating Godzilla tremendously in the form of OpenAI, Anthropic, Microsoft and Alphabet. Why do we not think that Anthropic would just buy a financial news site and build their own agentic Bloomberg for financial clients? They will have one of the largest IPOs of all time – they will have more money than seemed possible in 2023 to do whatever they need to do to continue these eye-watering valuations.
For years, I’ve heard media execs cluck their tongues that eventually the bottom will fall out for these AI firms. And yet they keep growing. Even Anthropic and OpenAI losing 50% of their market caps in the 12 months post their IPO stock-lock ups would still make them very very large companies.
We need to be much more vigilant.
I think one inevitable result is more mergers among media groups. Amalgamation of a giant mecha entity to take on AI firms isn’t a bad idea. I also see a lot of media groups, especially in B2B, going into full war mode with their own agentic research efforts, automated internal data mining, and new AI services and tools, especially for safely connecting their intelligence to these new agents inside the enterprise.
That’s a huge part of Miso’s most exciting work and I suspect B2B media will be in much better shape than B2C media within 2-3 years. I also think we’re underestimating the opportunity with this to sell to agents. But we need to think ourselves about what symbiosis between humans and AI looks like. B2B media is probably the best place to consider that, since professionals have workflows that we can really think about and build business value around.
As a result, I see a lot of shakeups as Google eats itself, and Godzilla steps on us in the streets. But I also see a lot of people building their own giant robots to punch big giant lizards in the face.
We’re excited to be on that side of the fight.
The Definitive AI Forum 2026 features 32 CEOs and industry leaders (including Lucky Gunasekara) from the following companies in the UK, US and EU:
