Insider’s Hack-AI-Thon

Insider had its ninth Hack Week — “Hack-AI-Thon” — in late March. Insider Product & Tech engineers worked on 23 projects centered around, you guessed it, artificial intelligence (AI).

The team asked for and received suggestions from throughout the company on how Insider might make use of emerging AI tools like ChatGPT to improve work efficiency, enhance internal communication, and discover innovative new ways to share stories with our audience.

Hack Week is a unique event that provides the Product & Tech team with the time and resources needed to step outside the box to brainstorm and collaborate on ideas to help the company “get better every day” — a core company mission.

The week is designed around specific goals, including:

• Building team member confidence and camaraderie.
• Unleashing creativity. The idea is to provide an opportunity to work on ideas that too often take a backseat to day-to-day priorities and workloads.
• Removing boundaries. A key component of innovation is to provide real flexibility in solving problems, without rigid borders.
• Fostering development. The week encourages team members to explore technologies with which they may not be familiar in their everyday work.
• Improving cross-team collaboration. Here, the goal is to foster a productive dialogue with the editorial and business sides of the company.

Here are some of the exciting projects that were developed during Hack-AI-Thon:

Tag-O-Matic 3000

Creative team: Chris Speziale, Corey Gahafer, Daniel Arita, Danny Feliz, Dasha Lary, Dawin Camilo, Emily Wong, Jenn Rourke, Jon Peck, Kelly Filush, Kunal Patel, Matt Maxbaum, Ronin Bennett-Vondersostensacken, Sarah Pai, Steffie Drucker

What It Does: CMS-integrated AI assistant that suggests category tags, recommends verticals, and helps journalists write context for their content.

This valuable tool emerged from a partnership between our product, tech, editorial, and SEO teams. It is intended to help journalists write context for their content and suggests category tags and verticals. It does this by using AI and SEO guidelines to allow editors to auto-generate vertical and category tags for their stories. A core underlying principle for the idea came from the Google principle for AI: be accountable to people. The core guideline is that it’s essential to ask users — here, editors — for permission. The creators also operated on the assumption that auto-generated responses are not necessarily better than what our talented and knowledgeable journalists come up with. As a result, the tool has an option for editors to regenerate (ie, come up with additional suggestions), add their own tags, or simply delete auto-generated recommendations. We are proud of this powerful new tool and expect that it will save our newsroom time, create more consistency around the use of verticals and category tags, and, ultimately, drive more traffic to our site.

Project Name: “I’m not listening”

Creator: Chase Gruber

What It Does: Use AI to listen to podcasts and summarize and/or allow us to ask questions about their audio content.

This hack project was inspired by partners in editorial, Jake Swearingen (Deputy Editor/Tech Analysis) and Walter Hickey (Deputy Editor/Data & Analysis). Many of our reporters listen to podcasts for inspiration for stories. However, listening to full podcasts can be time consuming and it’s hard to know when something is said that is of interest. The goal was to use A.I. to listen to and summarize the content of a podcast so reporters could quickly see whether something said in the podcast was of note. Once a podcast is selected (see screenshot below), behind-the-scenes, the program downloads the mp3, transcribes it, and summarizes it. In addition to getting a summary, an editor can ask a specific question about the content covered in the podcast. The tool can even provide timestamps so that the editor can go back and reference the provided answer.

Project Name: A dose of Inspiration ✨

Creator: Frederica Chen

What It Does: Helping editor’s overcome “writer’s block” with a dose of inspiration ✨.

Most if not all reports face writer’s block. So this program auto-generates possible article topics or product review ideas based on the author’s niche and past articles, along with current trends. After sharing the user data, A.I. generates possible topics or trending products to write about.

Hack week was an exciting event in which participants had the opportunity to explore and experiment with the latest advances in artificial intelligence technology.

The team showcased their skills and creativity by building innovative projects that pushed the boundaries of what is possible with AI and explored new ways to solve complex problems. They were able to learn from each other, receive feedback, and gain new insights into the capabilities and limitations of AI.

Hack-AI-Thon showed that this is just the beginning for using artificial intelligence to improve our work and product. It was especially exciting to craft a Hack Week around this exciting and emerging new tech — in fact, many of these new projects would not have been possible just six months ago.

The event also provided an extremely productive platform for developers to demonstrate our company values of being one team and being effective by collaborating with teammates on the editorial team to move towards Insider’s long-term mission of being the most beloved and influential journalism company in the world.

This blog post was co-authored by Thea Charles, Chief of Staff, Product & Tech and Joel Marino, Head of Learning and Development