We hear stories about the developing tech in AI. Many criticize AI for being biased towards specific things. But, they can be trained to help find the biased things and fight against them. One such application of the idea is being implemented by Primer. Primer is an AI startup which is using AI and Natural Language Processing to fight against the injustice done to several scientists. The Machine Learning algorithm made by them is used to scan the study and reports made by scientists present over the internet and give them recognition. Many startups like these use Wikipedia as their source but Primer has decided to fight against them. Through their algorithm, Primer is finding scientists whose works have been overlooked by Wikipedia. When the reports were studied, it was found that most of them were females who were not featured on Wikipedia.
With the reports Primer generated, they are aiming at showing their significant product to the world. In a post, John, the director of science at Primer said that “We have developed our own tool called QuickSilver. It reads around 500 million documents present over the internet. It identifies them and writes a draft article about the people and their work present in the document. Through this, the scientists who have their work published over the internet are recognized. For example, Primer showed one of their reports about a person named Teresa Woodruff.
She published her work on ‘Designing 3D-printed ovaries for mice’ but went unnoticed. Teresa was named as one of the Time’s Most Influential Persons in 2013 but sadly does not even have a Wikipedia page of her own. She has published her work on the internet and with all due credits, she deserves a page on Wikipedia.
Primer named their tool Quicksilver based on a tech book from Neal Stephenson titled, ‘Because we are nerds’. The tool has scanned nearly 500 million works and has identified 40,000 scientists who are not present on Wikipedia. It has generated a draft article on their work and about them as the result. Primer is also involved in three of Wikipedia editathons where the editors help themselves to improve Wikipedia.
When Primer was asked about their technology, they said that “Quicksilver is built on past works done by Google and some other companies. It uses Machine Learning to generate Wikipedia articles. We have started with working around Wikipedia but we have bigger goals. We want to change the online information ecosystem through our tool”.
Rather than just finding overlooked scientists and their works, there is more to the tool. Quicksilver is also being used to identify the pages which were edited a while ago. Many of these pages about people might miss some of their important works which they did later on.
However powerful Primer claims Quicksilver to be, it can be biased too. Primer uses the news stories to identify scientists hence it can affect the press. On this, Primer said that “We are trying to make this tool to assist the process. Our Machine Learning tool can be used to make the most important source of public information wider and updated”.