Would you be surprised if we tell you that a number of Artificial Intelligence labs are planning on using people’s voices to assess if they have COVID-19 or not? Because that’s what’s happening!
Currently, a research team comprising of members from Harvard and MIT is working on differentiating the voices of healthy people from those of people with the coronavirus. Their goal is to identify vocal signs that are only evident in the people carrying the disease.
Voca.ai (Israel-based startup company) is equipping the research team with audio samples. Alan Bekker, cofounder/CTO of Voca.ai revealed to Business Insider that the company in question launched a submissions portal. This portal welcomes people to submit their voice recordings to be used in the research. So far, the web portal boasts over 100 recordings from infected people and more than a thousand samples from people without the disease.
Even though the study in question has potential, we should remember that it is a new practice and the researchers working on it have zero to limited relevant experience.
Similar research is also going on in CyLab (Carnegie Mellon University). The lab in question also set up a raw AI tool that would assess people’s probability of being tested positive for COVID-19 using their audio recordings. Even though the lab put a disclaimer that the advice it would be giving wouldn’t be a genuine medical one, Carnegie Mellon called for the online test to be abandoned.
One of the researchers in CyLab, Rita Singh, understood the university’s concern with the online test. However, she believes that it’s high time for researchers to collect data for their study because once COVID-19 passes, it will (hopefully) never return.
Even Satrajit Ghosh, professor at MIT and Harvard, stressed on the need for more data to understand the diversity of symptoms as well as the changes the infected people undergo.
While the research in question is still in its early stages, it can be a true game-changer if everything goes right.
Read next: Apple & Google Are Ready For The Beta Testing of COVID-19 Tracking API
Currently, a research team comprising of members from Harvard and MIT is working on differentiating the voices of healthy people from those of people with the coronavirus. Their goal is to identify vocal signs that are only evident in the people carrying the disease.
Voca.ai (Israel-based startup company) is equipping the research team with audio samples. Alan Bekker, cofounder/CTO of Voca.ai revealed to Business Insider that the company in question launched a submissions portal. This portal welcomes people to submit their voice recordings to be used in the research. So far, the web portal boasts over 100 recordings from infected people and more than a thousand samples from people without the disease.
Even though the study in question has potential, we should remember that it is a new practice and the researchers working on it have zero to limited relevant experience.
Similar research is also going on in CyLab (Carnegie Mellon University). The lab in question also set up a raw AI tool that would assess people’s probability of being tested positive for COVID-19 using their audio recordings. Even though the lab put a disclaimer that the advice it would be giving wouldn’t be a genuine medical one, Carnegie Mellon called for the online test to be abandoned.
One of the researchers in CyLab, Rita Singh, understood the university’s concern with the online test. However, she believes that it’s high time for researchers to collect data for their study because once COVID-19 passes, it will (hopefully) never return.
Even Satrajit Ghosh, professor at MIT and Harvard, stressed on the need for more data to understand the diversity of symptoms as well as the changes the infected people undergo.
While the research in question is still in its early stages, it can be a true game-changer if everything goes right.
Read next: Apple & Google Are Ready For The Beta Testing of COVID-19 Tracking API
