Speakers Look At Artificial Intelligence For Sustainable Living

A panel of experts on artificial intelligence last week agreed that there are many opportunities for artificial intelligence to encourage a common good and sustainable living. And they set out several ways to get there.

The panel, “Artificial Intelligence for Common Good and Sustainable Living,” took place on 8 June.

The AI for Good Global Summit, which took place from 7-9 June, was organised by the UN International Telecommunication Union (ITU) and XPRIZE Foundation. The objective of the summit was to advance the development and democratization of artificial intelligence solutions which can address different global challenges, according to the organisers.

FeiFei Li, director of the Stanford Artificial Intelligence Lab, said it is “a very exciting time for artificial intelligence and computer vision.” Machine learning allows us to solve problems which have been difficult to solve in the past, followed Thomas Wiegand, executive director of the Fraunhofer Heinrich Hertz Institute and professor at the Technical University of Berlin.

Artificial intelligence needs to be imbedded in workflow and systems as we do not collect data for the sake of collecting data, said Antoine Blondeau, chairman of Sentient. The current challenge of artificial intelligence is to move from a narrow scale to a larger scale. The scaling of artificial intelligence is the start of the industrialisation process of artificial intelligence, Blondeau added.

As an example, artificial intelligence could be used to build a system in swimming pools to automatically detect drowning swimmers, Li said. There has not been a comprehensive technology that map out all the information that is collected by cameras and drones. Artificial intelligence may change this, according to Li.

Artificial Intelligence in Healthcare

Li explained how artificial intelligence can lead to improvements within the health sector and the workflow of healthcare.

Artificial intelligence can help to reduce the number of hospital-acquired infections, Li said. Hospital-acquired infection is a prevailing problem in the healthcare system, which can lead to fatal consequences for patients, she said. The “main driver” of hospital-acquired infection is hand hygiene, which is a hard problem to address, she added.

A recent advancement of computer vision technology allows the tracking of hand hygiene movements of medicines and nurses while ensuring the protection of their privacy. The system is very cheap and empowered by artificial intelligence collaboration. The accuracy of the technology is really high and better than most of the existing technologies, Li said. Hand hygiene is just the beginning, Lei said, adding that there are many other activities in healthcare where artificial intelligence can help with the continuous monitoring of both healthcare workers as patients.

Artificial Intelligence as Replacement for Census

The American Government spends billions of dollars each decade on conducting a census to obtain information about the American society, Li told the conference. She and her students tried to achieve the same results of a census by analysing available visual data.

Li and her students downloaded millions of pictures of Google Street View of the 200 most populated cities in America. The researchers detected all the cars in the pictures by using deep-learning systems and they obtained different kinds of information about the cars, such as the building year and model, by using another kind of deep-learning system.

The social make-up of the cities was mapped by using the information received from the analysis of the cars. There was a high correlation between the outcome of the census and the outcome of the research based on the pictures of Google Street View, she said.

Li used the data obtained from the cars to predict the outcome of the American presidential elections in 2008. The research revealed a high correlation between trucks and Republican voters at one hand and between highly environmental cars and Democratic voters on the other hand, Li said.

Li told the conference that this project illustrates the strength of visual data in our understanding of society.

Artificial Intelligence for All

Li is the co-founder and board member of AI4ALL, a non-profit group which aims at increasing diversity in the sector of artificial intelligence. Li underlined that there is a need to integrate people from diverse backgrounds. Results are “more creative” if a diverse group of people works together, she said, and it furthermore contributes to social justice.

Gender inequality in artificial intelligence has not received sufficient focus so far, according to Li.

 

Image Credits: Elise De Geyter