“`html
Information should guide every choice a contemporary organization makes. However, many companies face a significant oversight: They are unaware of what is occurring in their visual information.
Coactive is striving to transform this situation. The firm, established by Cody Coleman ’13, MEng ’15 and William Gaviria Rojas ’13, has developed a platform powered by artificial intelligence that can interpret data such as images, audio, and video to uncover fresh insights.
Coactive’s platform enables swift searching, organizing, and analyzing of unstructured visual material to assist businesses in making quicker, improved decisions.
“During the first major data revolution, companies enhanced their ability to derive value from their structured information,” says Coleman, alluding to data from tables and spreadsheets. “Yet currently, around 80 to 90 percent of the world’s data is unstructured. In the upcoming phase of big data, organizations will need to manage data like images, videos, and audio at scale, with AI playing a crucial role in unlocking that potential.”
Coactive is already collaborating with numerous large media and retail firms to help them comprehend their visual material without depending on manual sorting and tagging. This aids them in delivering the appropriate content to users more swiftly, removing inappropriate content from their platforms, and discovering how specific content affects user behavior.
More broadly, the founders assert that Coactive exemplifies how AI can enable individuals to operate more effectively and address new challenges.
“The term coactive signifies collaborative work happening simultaneously, and that’s our lofty aspiration: assisting humans and machines in working together,” Coleman states. “We believe this vision is increasingly vital because AI has the potential to either divide us or unite us. We intend for Coactive to act as a unifying agent that grants humanity a new set of superpowers.”
Granting vision to computers
Coleman encountered Gaviria Rojas in the summer preceding their first year through the MIT Interphase Edge program. Both would subsequently major in electrical engineering and computer science, undertaking projects like bringing MIT OpenCourseWare content to universities in Mexico, among others.
“That was a fantastic illustration of entrepreneurship,” Coleman reminisces about the OpenCourseWare project. “It was incredibly empowering to manage both the business and software development aspects. It inspired me to launch my own small web development ventures later on, along with taking the [MIT course] Founder’s Journey.”
Coleman initially delved into the capabilities of AI at MIT while serving as a graduate researcher with the Office of Digital Learning (now MIT Open Learning), where he employed machine learning to investigate how individuals learn on MITx, which offers massive open online courses designed by MIT faculty and educators.
“I was astounded that this transformative journey I experienced at MIT could be democratized through digital learning — and that AI and machine learning could be applied to craft adaptive systems that not only aid in grasping how humans learn but also provide more tailored educational experiences to individuals worldwide,” Coleman says regarding MITx. “That was also my first opportunity to work with video content and apply AI to it.”
Post-MIT, Coleman attended Stanford University for his PhD, where he focused on reducing obstacles to utilizing AI. This research enabled him to collaborate with firms such as Pinterest and Meta on AI and machine learning initiatives.
“There, I was able to glimpse into the future of what people aspired to do with AI and their content,” Coleman reflects. “I observed how leading firms harnessed AI to generate business value, and that’s where the initial inspiration for Coactive sparked. I wondered, ‘What if we developed an enterprise-grade operating system for content and multimodal AI to simplify that process?’”
Meanwhile, Gaviria Rojas relocated to the Bay Area in 2020 and began his role as a data scientist at eBay. As part of the relocation, he needed assistance transporting his sofa, and Coleman was the fortunate friend he reached out to.
“During the car ride, we recognized we both perceived an explosion occurring in data and AI,” Gaviria Rojas remarks. “At MIT, we were privileged to witness the big data revolution firsthand, where we observed individuals inventing technologies to unlock value from that data at scale. Cody and I realized there was another powder keg about to ignite, with enterprises amassing vast amounts of data, but this time it was multimodal data like images, videos, audio, and text. There was a missing technology needed to unlock it at scale. That was AI.”
The platform the founders ultimately created — which Coleman describes as an “AI operating system” — is model neutral, meaning the company can interchange the AI frameworks as models progress over time. Coactive’s platform features prebuilt applications that business clients can utilize for activities like searching their content, generating metadata, and conducting analytics to gain insights.
“Prior to AI, computers perceived the world through bytes, while humans perceived it through vision,” Coleman shares. “Now, with AI, machines can finally perceive the world as we do, and this shift will blur the lines between the digital and physical realms.”
Enhancing the human-computer interface
Reuters’ image database provides millions of photographs for journalists worldwide. Before Coactive’s involvement, the company depended on reporters to manually input tags for each image, ensuring the right photos appeared when journalists searched for specific topics.
“It was excruciatingly slow and costly to sift through all of these raw assets, resulting in people often neglecting to add tags,” Coleman states. “This meant that when you searched for items, the results were limited even if pertinent photos existed in the database.”
Now, when journalists on Reuters’ site opt for ‘Enable AI Search,’ Coactive can retrieve relevant content based on its AI system’s understanding of the specifics in each image and video.
“It significantly enhances the quality of results for reporters, enabling them to recount better, more accurate stories than ever before,” Coleman adds.
Reuters is not alone in facing challenges managing its content. Digital asset management constitutes a major component for various media and retail firms, which frequently resort to manually entered metadata for organizing and searching through their materials.
Another client of Coactive is Fandom, which is among the largest platforms globally for information about TV shows, video games, and movies, boasting over 300 million monthly active users. Fandom is employing Coactive to comprehend visual data in their online communities and assist in removing excessive violence and sexualized content.
“Previously, it took Fandom 24 to 48 hours to review each new piece of content,” Coleman explains. “Now, with Coactive, they’ve established their community guidelines and can generate more granular information in an average of about 500 milliseconds.”
In every application, the founders view Coactive as fostering a new paradigm in how humans collaborate with machines.
“Historically, in human-computer interaction, we’ve had to bend over a keyboard and mouse to input information in a manner that machines could interpret,” Coleman explains. “Now, for the first time, we can communicate naturally, sharing images and videos with AI, which can comprehend that content. This represents a fundamental transformation in our approach to human-computer interactions. The central vision of Coactive is that due to this change, we require a new operating system and a novel approach to working with content and AI.”
“`