高三英语完形填空分类训练：时事热点.docxVIP

完形填空 Artificial intelligence (AI) has amazing potential to change the world, and we’ve only just begun to scratch the surface. As AI matures and people move further away from distinct programming and monitoring of systems, unidentified bias (偏见) might make decisions continue for a long time that cause ___1___ harm for individuals and society. This bias might ___2___ input data or even the algorithms (算法) themselves. All too often, data sets are incomplete and the sample represented in the data set does not ___3___the population that the AI model is making predictions about—this is known as coverage bias. Some other types of bias related to input data include sampling bias, where data is not collected randomly from the target group, and participation bias, where users from certain groups ___4___ surveys at different rates than users from other groups. Still, another more challenging bias to identify is confirmation bias that occurs when a decision maker or analyst has a strong ___5___ belief or experience that affects their ability to consider alternatives. This could lead one to more strongly ___6___ data that confirms a preexisting belief. Bias resulting from AI algorithms themselves, or algorithmic bias, is equally ___7___. One example of algorithmic bias is implicit bias or unconscious bias, where data scientists ___8___ make associations or assumptions based on their mental models and memories that affect data modeling decisions. Implicit bias can ___9___ how data is collected and classified, or how systems are designed and developed. As machines learn, their conclusions and decisions affect people. Ethical (道德的) AI must understand these impacts and create governance and testing methods to ___10___ mistakes and inaccuracies. To create ethical AI, companies need to put the ___11___ of the individual at the center of data innovation. This means thinking about ___12___ rights as human rights and developing a comprehensive approach to data, including how we use AI.