During the 2016 U.S. elections, it became evident the extent to which fake news threatened democracy and had the power of shifting the truth in entirely different directions. Even though the spread of disinformation via digital culture was not so common a few years back, it’s clear now that it is one of the most significant issues that humanity is facing, given how much we rely on social media and technology.
The fake news problem is particularly difficult, given the lack of methods and tools to validate news and facts online. Therefore, TotalStories arose as a solution that could potentially fix this problem and change the world.
TotalStories is a social network that aims to democratize the production of news, by giving citizen journalists and the people who experienced the events first-hand the tools to share their experiences automatically and easily, and then use Artificial Intelligence (AI) to classify them and determine a quality index representing how unbiased and accurate they are. Readers automatically get informed about all the information they should be aware given their geographical position, interests, and many other factors, and are provided with automatic summaries and statistics showing them how relevant and valuable the articles are specifically for them.
Articles are automatically classified according to implicit elements such as tone, concepts, ideas, countries, and people discussed, among many other parameters. This way, a vast record of day-to-day information is being built on top of everyday news.
A Journey from Passion to Possibility
During the fake news crisis, some people proposed censorship as a solution. Edward Snowden, famous for having leaked classified information exposing global surveillance programs by the NSA, said that “the answer to bad speech is not censorship. The answer to bad speech is more speech.”
Snowden’s quote was powerful enough to inspire Diego Estevez, the founder of TotalStories, to create a platform that would help the people who experienced facts and news first-hand reach the readers directly. That way, he hoped to create an online hub for citizen journalism where news about virtually anything could be shared.
The project developed very fast: a minimum viable was made in early 2017 which showed that an ecosystem could successfully be formed around user-generated news and that traction in readers and professional writers could be achieved. Now, the platform counts with users from more than 45 countries and articles written in more than three languages, with writers ranging from startups in the San Francisco Bay Area to local newspapers in Argentina.
Meet the Tech-Savvy Leader
Diego Estevez is an 18-year-old entrepreneur and programmer who founded TotalStories when he was 15. Back in 2016, he became profoundly worried about the fake news problem. Even though he knew it would be an incredibly complicated problem to solve, he believed he could contribute by applying his passion for computer science and startups to build a new way of writing and sharing news. His vision was initially helping isolated communities and citizen journalists to reach a broader audience while creating an online news repository of events, concepts, cities, and people. One of the fundamental philosophical values behind the project was simply the democratization of news writing and the empowerment of the people who lived the facts first-hand to share them without a middle-man — to some extent, creating a YouTube for news.
TotalStories was his second startup though, having founded when he was 14 years old a videogame server that went viral and was making nearly $1000 in profits weekly. Additionally, he’s also the founder of PanamaHackea, an NGO that seeks to bring entrepreneurship and the hacker culture to Panama through hackathons, events, and providing seed-funding to startups. He briefly worked as a data scientist at the San Francisco company Grandata and has done research in consensus and Proof of Stake algorithms with the hope of making the blockchain’s scale better. He has spoken at TEDx twice; first about the potential of using low-cost and do-it-yourself technology to connect isolated communities around the world, and the second time about the importance of following your ideas as a teenager.
An Ambitious Handout to AI
TotalStories is an experiment-turned-into-startup project that seeks to show that a reliable news ecosystem can be built without censorship, focusing instead on user-generated content and artificial intelligence.
On the one hand, for users it provides a much more improved way of reading news, automatically getting summaries and analysis specifically determining how valuable the articles are while offering a suite of statistics showing how the report could be biased and what the writer is trying to convey. Given that the pieces of writing are automatically analyzed for many tacit elements such as sentiment, concepts, places, and ideas, it allows the system to classify the articles into very specific categories. From these, they could later be recommended and shared to users that the algorithm predicted they should know about.
On the other hand, for writers, it automatizes the process of sharing news while enriching those with related stories, events, people, and topics. Simultaneously, it provides a standard of quality from which to improve the reports (given that they receive a quality index), which incentivizes writers to provide the platform with direct links to sources and evidence. Writers also get a reputation index based on their articles, which then becomes an unbiased way of ranking and comparing newspapers by “trustworthiness”.
An Innovative Avenue of Reportage
TotalStories, particularly, showed a very innovative application of AI, proving that it could successfully be used throughout the entire cycle of news publishing: from the writing of the article to when the readers consume it. Such a degree of automatization seems to have successfully built a bridge between the citizens who experience facts and the consumers, in a trustworthy and efficient manner.
Boosting Technology Engagements
Diego believes that IA is enabling entirely new ways of interacting with data and computers, potentially creating new ways of interacting among users and consumers. “Given the availability of large volumes of data and better ways of classifying it, more specific day-to-day solutions and potential business ideas are arising as a result of better personalization techniques and the use of algorithms to successfully generate content that could look human-generated,” Estevez said.
For networks based on human-generated content like TotalStories, Diego thinks that AI could be used to lower the point of critical mass and to create a much smoother experience for consumers, by using procedurally generated content to fill any gaps in the human-generated one.
“Simultaneously, given that the number of data increases by the day, I think that personalization will also become a vital aspect for many businesses and prove an extremely useful niche for startups,” he added.
AI: Suiting Up for Present, Setting Benchmarks for Future
Diego says it’s interesting to see how new applications are being explored and how now more than ever hobbyists and average citizens are starting to develop solutions on their own. “Tools like IBM Watson and DialogFlow are democratizing the access to artificial intelligence development, which is connecting more day-to-day elements and problems to AI-powered solutions. It’s clear that chatbots and voice-controlled devices are becoming common and incorporated into the day-to-day life, and it’s awesome that some of the developers are teenagers or people who never had the chance of developing an IA-powered solution before”.
Creating a Hub for Citizen Journalism
Even though the TotalStories’ platform is still under experimentation, users that have joined the social network range from average-citizens to high-school newspapers writing their blog posts and weekly news. “These people are collaborating to create a very rich archive of news and a better way of storing and classifying day-to-day information,” he said. News articles have been classified according to more than 200 implicit topics, and many of these have been exclusive to TotalStories.
Winning Every Challenging Horizon
Commenting on the challenges Diego said, “From an engineering perspective, it was hard to develop a system that could scale to a high-rate of news production with a real-time AI analysis. The latter was certainly not easy to develop either and involved making many ethical choices on what should be considered biased and unbiased. Without a doubt in mind, the algorithms that rank the articles and create the quality indexes also took a significant amount of work. At many points, I felt overwhelmed by their complexity, but luckily the fulfillment that I received when I saw direct evidence that they had the potential of contributing to solving the fake news problem and that the algorithms were actually working, was very motivating. It turns out that they ended up becoming a hybrid of several technologies, ranging from APIs to our own Scikit models”.
Estevez also commented that “It was very tough getting solid based of readers at the beginning though. Even though it seems obvious now, I realized that if I had no clear way of getting the readers to stay, then it would be really hard to get newspapers to write exclusively on the site. That’s when I realized that I could populate the site using RSS to make up for the lack of mainstream articles.”
Envisioning the Future Roadmap
Diego thinks there will be a tendency to decentralize news even more. He firmly believes that the blockchain can provide a much more transparent way of storing and analyzing a wider variety of data. He places special emphasis on the fact that since they are publicly stored, they could lead to better interactions among different services and better applications or inferences from the data. “I think that the blockchain also promotes the idea of having a direct path between the readers and the citizen journalists. The fact that information is immutably stored means that no censorship could take place either. It’s also interesting to see how new business models are being created out of the intersection between artificial intelligence and this new decentralized technology. There’s certainly a lot of potential for the values on which TotalStories was built to be implemented and further developed using them, ” he said.
“I think that all of these technologies could lead to higher volumes of high-quality data being generated on top of TotalStories. Not only would this lead to a much heavier reliance on automatically generated articles based on isolated information found from different sources, but more and better ways of getting the reader involved on what used to happen behind the scenes,” he mentioned.