人工智能一张照片可判断性向
<p><span word="Artificial">Artificial</span> <span word="intelligence">intelligence</span> <span word="can">can</span> <span word="accurately">accurately</span> <span word="guess">guess</span> <span word="whether">whether</span> <span word="people">people</span> <span word="are">are</span> <span word="gay">gay</span> <span word="or">or</span> <span word="straight">straight</span> <span word="based">based</span> <span word="on">on</span> <span word="photos">photos</span> <span word="of">of</span> <span word="their">their</span> <span word="faces">faces</span>, <span word="according">according</span> <span word="to">to</span> <span word="new">new</span> <span word="research">research</span> <span word="suggesting">suggesting</span> <span word="that">that</span> <span word="machines">machines</span> <span word="can">can</span> <span word="have">have</span> <span word="significantly">significantly</span> <span word="better">better</span> "<span word="gaydar">gaydar</span>" <span word="than">than</span> <span word="humans">humans</span>.</p><p>一项新研究显示，人工智能可以通过人脸照片精确识别出这个人是直男还是同性恋，该研究认为，机器的“<span word="gay">gay</span>达”（同志雷达）比人类准确得多。</p><p><span word="The">The</span> <span word="study">study</span> <span word="from">from</span> <span word="Stanford">Stanford</span> <span word="University">University</span> – <span word="which">which</span> <span word="found">found</span> <span word="that">that</span> <span word="a">a</span> <span word="computer">computer</span> <span word="algorithm">algorithm</span> <span word="could">could</span> <span word="correctly">correctly</span> <span word="distinguish">distinguish</span> <span word="between">between</span> <span word="gay">gay</span> <span word="and">and</span> <span word="straight">straight</span> <span word="men">men</span> 81% <span word="of">of</span> <span word="the">the</span> <span word="time">time</span>, <span word="and">and</span> 74% <span word="for">for</span> <span word="women">women</span> – <span word="has">has</span> <span word="raised">raised</span> <span word="questions">questions</span> <span word="about">about</span> <span word="the">the</span> <span word="biological">biological</span> <span word="origins">origins</span> <span word="of">of</span> <span word="sexual">sexual</span> <span word="orientation">orientation</span>, <span word="the">the</span> <span word="ethics">ethics</span> <span word="of">of</span> <span word="facial">facial</span>-<span word="detection">detection</span> <span word="technology">technology</span> <span word="and">and</span> <span word="the">the</span> <span word="potential">potential</span> <span word="for">for</span> <span word="this">this</span> <span word="kind">kind</span> <span word="of">of</span> <span word="software">software</span> <span word="to">to</span> <span word="violate">violate</span> <span word="people">people</span>'<span word="s">s</span> <span word="privacy">privacy</span> <span word="or">or</span> <span word="be">be</span> <span word="abused">abused</span> <span word="for">for</span> <span word="anti">anti</span>-<span word="LGBT">LGBT</span> <span word="purposes">purposes</span>.</p><p>这项斯坦福大学的研究发现，计算机算法能正确区分直男与同性恋，准确率高达81%，对女性性取向判别的准确率为74%。这一研究引发了人们对性向的生物学起源、人脸识别科技的道德伦理以及此类软件对个人隐私可能造成的侵犯，或被滥用于反同性恋、双性恋及变性人群体等问题的争议。</p>
页:
[1]