Most investors won't even consider a startup pitch unless they know the founders or have had a "warm intro" to them.
But 25-year-old Dhruv Gulati may be the exception to the rule, after he emailed billionaire investor Mark Cuban without any introduction to pitch his anti-fake-news startup Factmata — and persuaded Cuban to invest.
Cuban has now participated in its seed round, along with Zynga founder Mark Pincus and Brightmail founder Sunil Paul.
"I had a key list of people I wanted to have on board," Ghulati told Business Insider, adding that Cuban was among them. "Every single investor we have has some element of being extremely strategic and important for the business. It was about making that clear to the investor, as opposed to just, 'Why am I randomly emailing you?'"
Factmata wants to use artificial intelligence to tackle the proliferation of fake and misleading news. The startup aims to be a cross between Wikipedia and Quora, with a community of users fact-checking or marking news articles for quality with the help of AI. Those users may include everyday internet users but also journalists.
The startup is building its first product for launch next year: a news aggregator designed to show a quality score and offer extra links for context. Eventually, Factmata may offer its underlying technology to public-relations firms, media outlets, and other organisations for a fee.
Ghulati's thesis is that journalism's prevalent ad-funded model has encouraged news outlets to pursue eyeballs through sensationalism and misinformation. The rise of that model online coincides with the decline of the subeditor in newsrooms, the people who rigorously check articles before they are published.
All of this, he said, has helped contribute to waning trust in the news media — which in turn may have led to the rise of fake news.
A journalist may be skeptical that any kind of quality-checking on articles can be outsourced to machine learning. But Ghulati said it was about building tools to assist, not replace, the fact-checking process.
He said: "Not all of it can be done with machine learning. Hopefully there'll also be input gathered from a big community of users, the Wikipedia model. That's what we're aiming for, a credibility score on information."
The team has a strong pedigree. Its chief technology officer, Robert Stojnic, was formerly a developer at Wikipedia, building out the platform's search function. And Ghulati's cofounders are two machine-learning specialists: UCL researcher Sebastian Riedel and UCL research associate Andreas Vlachos.
"I was impressed by the team's pedigree, technical talent, and sheer drive to solve this problem," Cuban said in a statement.
"If we want to solve fake news, thinking about it at web scale via artificial intelligence and automation is the only way. And being outside the media or fact checking world allows them to see the problem in a different way. Factmata is a group of entrepreneurs trying to solve a challenging problem with an amazing mission."
Factmata isn't the only project trying to tackle fake news. Facebook has added features designed to alert people to misinformation, while Wikipedia founder Jimmy Wales launched the community-powered news outlet Wikitribune earlier this year.
Ghulati is sceptical that Facebook can solve the fake-news problem by itself.
"We're a platform fully focused on quality of information," he said. "With Facebook, it's not in their model, it's not what they think about. And they're ads-incentivised. Clearly there's a demand from the public to solve these problems, and whether platforms can do that to an adequate level — who knows? But we're 100% focused on it."
This article originally appeared on Business Insider.