1. One thing that helps
Rarely do we get the opportunity to directly help technology solve a major issue. Child exploitation on the internet is something that technology has enabled, but it’s also something that technology can fight – with your help.
My Pictures Matter is inviting adults to share childhood photos of themselves. These will be used to help train their machine-learning project to identify children in photos, and detect and classify child exploitation material. The approach is purely voluntary and transparent in the way they handle your data. Everyone in the photos must now be over 18 and agree to their photos being used. They are not accepting photos with any nudity – not even that embarrassing one of you in the bath with your sister!
With more than 33,000 reports of online sexual child exploitation in Australia alone, this community-driven approach will be of great help in fighting against such crime. Currently, officers investigating cases have the horrible job of examining images for extended hours. This work is painstaking, inefficient and can cause psychological trauma. Hopefully, this new approach will help law enforcers rescue victims of exploitation and shut down offenders much faster.
2. One to be wary of
Oh my god! Back in 2018, Microsoft and the province of Salta in Argentina apparently deployed an algorithm that predicts which low-income “future teens” would be likely to get pregnant.
A survey was conducted with 19,000 girls aged from 10-19, mostly from migrant families, where they provided data such as age, country of origin, disabilities and whether their house had hot water. The data was then fed to an algorithm called Technology Platform for Social Intervention (not Orwellian at all). It then supposedly used this data to predict which of the participants are likely to get pregnant in the future.
The fact that it was Microsoft running this project with no transparency, in a region notorious for surveillance and population control, is pretty shocking. The girls who were surveyed mostly came from disadvantaged backgrounds. Most worryingly, there were even “territorial agents” used to take photos and GPS locations of the girls who were likely to get pregnant according to the algorithm. The project seems to have been conducted with no concern for how it may affect those being studied.
Technology can be amazing or scary. Here is a clear demonstration that it is the people deciding which it will be.
3. One to amaze
Cornell University researchers are helping robots sweat like us!
When we get hot, we sweat to cool down. When metal-based machines get hot, they require internal fans to cool down. Soft robots, made from flexible synthetic materials, also have inherent heating problems due to their materials that retain heat rather than dissipate it.
The researchers mimicked sweat with a hydrogel material that squeezes water when it reaches a certain temperature. This sweat helps the robot operate longer without overheating.
Robotic sweating (what a term) can cool the robot lower than the ambient temperature. Fans just circulate air so the cooling effect is limited by the environment’s ambient temperature. Ever tried sitting in front of a fan that blows hot air? With sweat, robots can cool off anytime they reach around 30 degrees Celsius. The temperature threshold can also be adjusted as required.
In addition to temperature control, the ability to secrete liquid can be used to reduce friction allowing soft robots to crawl more efficiently. The technique can also be used to imitate the process of digestion and absorption in order to clean a contaminated area.
This technique still needs more progress as there remains the question of how to replenish the sweat once it’s used up. Maybe we will give robots the ability to drink to replenish, and maybe a can of deodorant if they can find their underarms!