In 2010, Karthik Dinakar and Birago Jones, students at the Media Lab, teamed up to create a tool designed to aid content moderation teams at companies including Twitter and YouTube. The project generated widespread interest and soon they were invited to the White House to demonstrate their technology designed to identify concerning posts on these platforms. While fine-tuning their demo, a significant problem emerged: the model was unable to understand posts written in teenage slang and other indirect language. Unlike conventional machine-learning models, Dinakar and Jones needed a method which could understand these posts which led them to develop point-and-click tools allowing non-experts to build machine-learning models. They named their creation “Pienso.”
This technology doesn’t simply allow non-experts to use existing models; instead, Pienso empowers them to build their large language models capable of detecting misinformation, human trafficking, weapons sales and more. While refining their software, they worked with local students in Cambridge, Massachusetts, who provided the necessary input to effectively train the models. Dinakar and Jones realised that empowering domain experts in the field of AI was vital for its success.
Their project came to fruition while both Jones and Dinakar were graduate students at the MIT Media Lab. Following their Master’s degrees, the pair began working on Pienso part time until 2016 when Dinakar completed his PhD at MIT, coinciding with a surge in the popularity of deep learning. The Industrial Liaison Program and the Startup Accelerator at MIT have been instrumental in connecting the founders with early partners.
An early partner, SkyUK, utilized Pienso to understand common customer issues, integrating these models into their system to handle half a million customer calls every day. According to Dinakar and Jones, this has saved SkyUK over £7 million by shortening call times.
In 2020, amidst the Covid-19 outbreak, government officials sought the help of Pienso to analyse thousands of research articles on coronaviruses. Information retrieved from this helped the government identify and strengthen critical supply chains for essential drugs.
Pienso helps businesses compile and analyse vast amounts of data without resorting to other AI companies. It’s compatible with both internal servers and cloud infrastructure, which, combined with its user-friendly interface, makes it a preferred choice for businesses. It has also partnered with GraphCore, a machine-learning computing platform to further increase speed and efficiency.
The founders believe Pienso is ushering in a future where effective AI models are built by those who best understand the problems they aim to solve, rather than simply using an existing one-size-fits-all model. This allows for a more nuanced and effective approach to problem-solving, utilising the knowledge of experts in the field.
