Artifact: The AI-Powered News App by Instagram Founders

Artifact, a news discovery app developed by Kevin Systrom and Mike Krieger, the co-founders of Instagram, is a testament to the recent advancements in artificial intelligence (AI). While the concept of Artifact predates the rise of AI language models like GPT (Generative Pre-trained Transformer), Systrom asserts that the app emerged as a result of their desire to leverage the capabilities of machine learning to address a significant problem.

The problem Artifact seeks to solve is the challenge of finding personally relevant and high-quality news articles amid an overwhelming sea of clickbait, biased content, and low-value distractions. Rather than relying on suggestions from others, Artifact's selection and ranking of news articles are primarily determined by the content itself. The app aims to provide users with the news stories they genuinely want to see, sourced from reliable publications.

To accomplish this, Artifact's small team has committed to the AI revolution. The app employs a proprietary large language model that represents news articles as numerical hashes, enabling efficient matching of user preferences based on their clicks, reading time, and stated interests. Systrom emphasizes that Artifact is not focused on generating text like GPT or Bard but on understanding it.

While Artifact doesn't currently engage in conversations with users like ChatGPT, it does collaborate with OpenAI, granting access to the GPT-4 API. Users can choose to have news articles summarized into concise bullet points using AI-generated summaries before committing to reading them, although Artifact warns that these summaries may contain errors.

In its latest endeavor, Artifact aims to tackle the issue of clickbait headlines. Users can flag clickbait stories, but the app also allows them to report problematic headlines. In such cases, the content is submitted to GPT-4 for analysis, which then generates a more descriptive headline to replace the original one. While Artifact currently shares the AI-generated headline only with the user who reported it, if multiple users flag the same clickbait title, the AI-generated headline replaces the publisher-provided one for all Artifact users. Systrom envisions a future where the system can autonomously identify and replace misleading headlines without user intervention.

While Artifact's primary goal is to deliver relevant news to users, Systrom acknowledges that the app's mission may extend beyond journalism. He draws parallels to the evolution of other successful companies, such as Apple with personal computers, Amazon with bookselling, and Facebook with college networks. Systrom's overarching vision is to enable people to consume what matters most to them, leveraging the power of machine learning as the driving force behind the next wave of innovation.

In reflecting on Systrom and Krieger's experience with Instagram, it becomes evident that Artifact's trajectory could similarly diverge from its initial focus. The transformative power of AI has the potential to reshape the educational landscape, prompting a reconsideration of how writing skills and critical thinking are valued in an era where machines can generate content. Nonetheless, Systrom underscores the importance of nurturing human intellectual abilities and ensuring that students develop the skills to think independently, while also embracing the partnership between humans and AI.

In conclusion, Artifact, conceived by the Instagram founders, harnesses AI to curate personalized news feeds and combat clickbait. The app leverages its own language model to understand content and employs the GPT-4 API for summarizing articles and rewriting misleading headlines. Systrom envisions Artifact expanding its scope beyond news consumption, driven by the transformative potential of machine learning.