algorithms are used to find the quickest and cheapest route home to entice us to buy things we don't necessarily need but may want or to keep us watching so what is an algorithm the algorithms perform step-by-step calculation according to the the formula or model and then you get a certain outcomes poorly thought out algorithms can lead to dangerous or unintended outcomes the robo debt scheme for example handed out hundreds of thousands of incorrect debt notices meaning some people who didn't owe anything were being told they did so we can actually track professor fang chen says
algorithms shouldn't be feared even if they are being used to make recommendations all the time some of them has been embedded as the recommendation system those ones this happened at the background but most of the decisions nowadays still have a level of human intervention during the covert crisis professor chen's team has been using algorithms to look at how covert 19 is impacting local government areas including rental prices community sentiment and to determine risk levels for the homeless and aged care in certain areas the algorithms are calculating which regions change the most which region is less
and then feedback those information to government and then for decision making because humans make the codes algorithms can result in biased or racist outcomes in the uk they've backed down from a decision to use grades determined by an algorithm that used a school's previous performance rather than their teachers individual marks many kids in state schools or from underprivileged backgrounds saw their marks dramatically reduced in the u.s there are concerns about computer-generated risk assessments to help judges decide if people should be bailed how you design the ai systems that make a huge difference and particularly is
what sort of a feeds you have to the ai system ai is like a child and what you fit into ai what do you teach him or her and that's what sort of outcome is going to get professor chen says being transparent and conscious of the data as well as testing as much as possible will help so what will our algorithmic future hold do i see in future algorithms make decisions themselves yes however i firmly believe that those processes need to be validated through a human when those algorithms make the same decision as we do
or make a better decision than we do then that's the time you fully just let it automatically run you
