In a recent SafetyDetectives interview, Arjun Narayan, Head of Global Trust & Safety at SmartNews, shared insights into his role and the company’s mission. Leading the team responsible for platform safety and integrity, Narayan highlighted SmartNews’s commitment to providing diverse, high-quality news content that avoids creating filter bubbles. He also discussed the company’s proactive measures against misinformation, challenges in using machine learning for news curation, and the company’s strong focus on user privacy and publisher relationships.
Can you talk about your journey and your role at SmartNews?
I head the Trust, Safety, and Operations excellence team at SmartNews. My team is mandated with keeping our platform safe and authentic for our users while ensuring Platform integrity. I joined SmartNews just as the world had just about recovered from COVID and things were normalizing. The SmartNews mission focuses on providing high quality, authentic, and diverse news to combat media distrust and reduce polarized echo-chambers, while driving quality readership to publishers – this is what initially drew me to the platform.
Since establishing the trust & safety team, we have strived to ensure that our algorithm and recommendation engine is optimizing for the highest quality content signals, auditing for bias, removing any policy non-compliant content, and last but not the least, identifying and removing low quality content.
With an increasing number of news aggregation platforms emerging, what sets SmartNews apart from its competitors?
We pride ourselves in providing our users a large selection of high quality and diverse news content. Our algorithms deliver an impartial and diverse selection of the best articles from a wide variety of sources from around the world. Many news apps focus on personalizing the news based on your social network and favorite topics, but this often creates a ‘filter bubble’ effect where you read the same thing over and over again. This gets boring, not to mention amplifies personal biases.
We recently launched SmartTake, an opportunity to experience newsfeeds in a new, unique way that balances the day’s news with a dash of mindfulness and uplifting stories. This as opposed to some of the more commercially successful news platforms that optimize for doom scrolling.
In an age where misinformation spreads rapidly, what proactive measures does SmartNews take to combat fake news?
We employ a multifaceted approach to combat fake news. This includes assessing the credibility of news sources, verifying information consistency across diverse platforms, collaborating with fact-checking agencies, and fostering an engaged community that actively provides feedback and flags questionable content. Additionally, we recognize the significance of media literacy in countering misinformation, leading us to introduce the “News From All Sides” feature a few years ago. This feature, situated under the politics tab in the SmartNews app, enables users to explore articles on a specific news event categorized along a spectrum from the most liberal to the most conservative. Our aim is to enhance transparency, expose readers to diverse perspectives, and make the discovery of varied content a delightful and serendipitous experience.
Can you talk about the challenges of using machine learning to curate news while avoiding reinforcing readers’ existing beliefs?
Employing machine learning for news curation introduces challenges centered around avoiding algorithmic bias and filter bubbles. The risk of reinforcing users’ existing beliefs is heightened when algorithms are trained on biased data, potentially creating echo chambers. Algorithms emphasizing user engagement may prioritize sensational content, skewing the presentation of diverse perspectives. The dynamic nature of news and ethical concerns further complicate the development of algorithms that provide balanced and unbiased news recommendations.
At SmartNews we understand this very clearly, in fact SmartNews was founded to solve this very problem. Addressing these challenges necessitates a commitment to fairness, transparency, diverse training datasets, and ongoing refinement of algorithms to ensure a more inclusive and reliable news consumption experience. The push and pull between discovery, diversity, and relevance are all inputs into our ever-evolving algorithm.
How does SmartNews handle user data, especially regarding their reading habits and preferences?
Respecting and protecting the privacy of our users is of utmost importance to SmartNews. We regularly update and refresh our privacy policy to explain our practices regarding the collection, use, and disclosure of information that we receive when you use our Services. You can read our privacy policy at
https://www.smartnews.com/privacy/
How do you navigate and manage relationships with content creators and providers, especially around issues like content sharing or copyright?
SmartNews has a unique publisher-friendly model: We help publishers reach, retain, and monetize new mobile audiences through our SmartView First program. Over the years, we have developed a deep and trusting relationship with our partner publishers and hope to continue to retain and further build on this trusted relationship. As far as copyright; SmartNews respects the intellectual property rights of our publishers and this is reflected in our copyright policy. You can access our copyright policy as well as copyright owners can notify us of any infringement at https://www.smartnews.com/copyright/.