In 2010, Media Lab students Karthik Dinakar SM ’12, PhD ’17, and Birago Jones SM ’12 partnered for a class project aimed at developing a tool to assist content moderation teams at firms such as Twitter (now X) and YouTube. The widely acclaimed project led to an invitation to demonstrate their innovation at a cyberbullying summit at the White House. After facing a crucial issue that nearly derailed the project before the White House event, the students realized that the lack of understanding of the type of data they were dealing with was the problem, not the model itself. They thus concluded that the task of creating such models was not solely for machine-learning engineers but individuals who best understand the relevant data.
This understanding inspired the duo to create point-and-click tools that enable nonexperts to create machine-learning models. The tools formed the foundation of Pienso, a platform that facilitates the building of large language models for detecting misinformation, human trafficking, weapons sales, and more, without writing any code. Pienso presented the system at local schools in Cambridge, Massachusetts, where students were allowed to train the models. This collaboration yielded models that were superior to anything that the software could have come up with on its own.
Pienso’s founders reported that the system has been applied to a wide range of tasks. Early on, telecom giant SkyUK collaborated with Pienso to develop models to better understand its customers’ common issues. These models currently support the processing of approximately 500,000 customer calls daily, which the founders claim has saved the company over £7 million by reducing the length of service calls.
In 2020, government officials sought help from Pienso to better understand the novel coronavirus. The tool enabled experts in the field of virology and infectious diseases to establish machine-learning models to analyze thousands of research articles about coronaviruses. The work was said to have assisted in identifying and strengthening critical drug supply chains, such as the widely used antiviral drug remdesivir.
Pienso’s platform offers a beneficial alternative for businesses that are compelled to give up their data to use services from other AI companies. Companies can import data with ease, prepare it for deep learning, analyze it, and structure it accordingly, all without writing a line of code. This year, the company announced its alliance with GraphCore, which offers a faster and more efficient computing platform for machine learning.
The founders of Pienso are focused on driving a future where more efficient AI models are created to address unique problems by individuals who possess a deep understanding of the problems they are working to solve. They highlighted the importance of bringing together different models to work collaboratively, orchestrated by those most knowledgeable about the relevant data.
