
With the election approaching, fake news is circulating online. But this year there’s a new threat: AI-generated deepfakes. Deepfakes can show politicians saying words they’ve never said. They can also show celebrities making spurious candidate endorsements. The Investigation Bureau says it will take action against deepfakes ahead of the 2024 general election.
The bureau has software tools for detecting deepfakes and removing them from the web within 48 hours. It’s also ramping up public education on how to identify deepfakes. To spot a fake video, watch out for eyes that don’t blink, or blink in a way that seems unnatural. Another sign of a deepfake is a misalignment between gaze direction and orientation of the face. Besides these traits, you can also watch out for blurriness along the edges of the face, or distortions of the teeth, glasses, or jewelry.
For more Taiwan news, tune in:
Sun to Fri at 9:30 pm on Channel 152
Tue to Sat at 1 am on Channel 53
大選步步逼近,網路上假消息越來越多,但今年還有新的威脅,AI變臉技術發達,就怕有心人偽造名人拉票,法務部調查局嚴陣以待,加強宣導民眾辨識,也備好識別軟體,48小時內下架假影片。
調查局表示,一般民眾也可以透過肉眼識別,眨眼的頻率自不自然,視線方向與臉的方向一不一致等,除了這些基本的生理特徵外,也可觀察臉部邊緣是否模糊或出現色差,或是牙齒、眼鏡、飾品出現瑕疵等。
更多新聞內容,請鎖定:
民視台灣台(152頻道)週日至週五晚上9:30
民視新聞台(53頻道)週二至週六凌晨1:00
Related News