메뉴
BL
Wired AI • 33일 전

교단에 삶을 바친 교사들, 그리고 시작된 딥페이크의 악몽

IMP
7/10
핵심 요약

WIRED는 학생들이 생성형 AI로 만든 성적 허위 이미지(딥페이크)로 피해를 입은 여러 미국 교사들의 사례를 보도했습니다. 캘리포니아의 대리교사 루이스 데산티아고는 자신의 사진이 변조된 부적절한 이미지가 소셜미디어에 확산되며 두 학교에 걸친 괴로움을 겪었고, 안전을 느끼지 못해 근무 지역으로 복귀하지 않기로 했습니다. 이는 AI 이미지 생성 기술이 학교 현장에서 교사와 학생 모두에게 심각한 피해를 주고 있으며, 'Take It Down Act' 같은 법적 대응의 중요성을 보여줍니다.

번역된 본문

캘리포니아주 로스바노스의 대리교사 루이스 데산티아고에게는 모든 것이 몇 통의 문자 메시지에서 시작되었다. 3월 12일, 그는 다른 파트타임 일자리인 커피숍에서 일하던 중 휴대폰이 켜지는 것을 보았다. 동료 바리스타가 소셜미디어에서 AI로 만든 사진 하나가 돌아다니고 있다고 했다. 그녀의 사촌이 그것이 데산티아고인지 확인하고 싶어 했다.

처음에 데산티아고는 걱정하지 않았다. 부적절한 사진은 틱톡(TikTok)의 커뮤니티 가이드라인을 위반하므로 대부분 빠르게 삭제되리라 생각했기 때문이다. "그 이미지를 마침내 봤을 때, 배를 한 대 맞은 것 같았어요"라고 데산티아고는 WIRED에 말했다. 이미지의 일부는 대학 시절 직장에서 찍었던 거울 셀카와 비슷했다. 하지만 작업복 대신 속옷을 입고 있었고, 원본 사진처럼 평범하게 서 있는 것이 아니라 선정적인 포즈를 취하고 있었다.

이 문자들은 두 학교에 걸쳐 학생과 행정관리자까지 연루된 수개월간의 고난의 시작이었다. 그리고 데산티아고는 이런 방식으로 피해를 입은 유일한 교사가 결코 아니다. WIRED는 학생들이 자신들에 대해 아마도 AI를 사용해 성적인 사진과 영상을 만들었다고 주장하는 세 명의 다른 교사들과 인터뷰했다. 그들은 이 사건들이 깊이 충격적이고 실망스러웠으며, 어떤 경우에는 자신의 일과 학생들과의 관계를 바라보는 방식을 다시 생각하게 만들었다고 말한다.

"다른 학생들도 포함될 수 있다"

문자를 받은 후 데산티아고는 이미지 뒤에 누가 있었는지 추론하기 시작했다. 그것을 공유한 계정 중 하나는 틱톡 프로필에 실명을 사용하고 있었는데, 그가 정기적으로 대리 근무를 하던 지역 중학교에서 알던 학생이었다. 데산티아고는 그 계정을 팔로우하고 상호작용한 사람들과 학생 친구 그룹에 대한 자신의 지식을 바탕으로 다른 두 학생의 관여를 의심했다. 그는 바로 다음 날 그 중학교에서 근무할 예정이었다. 데산티아고는 이 사건을 교감에게 보고했고, 교감은 학교가 '처리하겠다'고 확실히 말했다.

그날 저녁 늦게, 그는 틱톡에서 자신의 이름을 검색해 그 이미지를 찾아보기로 했다. 그러던 중 다른 학생이 자신을 사칭하는 계정을 만든 것을 발견했다. 그 계정은 그의 이름을 약간 다르게 철자해 사용했고, 그 사진을 프로필 사진으로 쓰고 있었다.

데산티아고는 이후 몇 주 동안 학교 행정관리자들, 학생들의 학부모, 심지어 경찰과도 왔다 갔다 했다고 말한다. "이 학생들이 저에 대한 이런 이미지를 만들고 게시하는 것을 두려워하지 않는다면, 문자 메시지와 단체 채팅을 통해 개인적으로 어떤 것을 유통하고 있을지 걱정됩니다"라고 데산티아고는 교장에게 썼다. "이것이 다른 학생들도 포함할 수 있다는 점을 유의하시기 바랍니다."

이미지를 만든 것을 인정한 한 학생은 교내 정학 징계를 받았다. 데산티아고는 나중에 그 이미지가 스냅챗(Snapchat)에도 공유되었다는 말을 들었다. 하지만 그는 이미지가 계속 확산되었고 다른 학생들이 게시하기 시작했다고 말한다. 또한 자신의 틱톡 계정에 여러 건의 저속하고 성적인 댓글을 받았다고 한다. 데산티아고는 관리자들에게 "이 문제가 완전히 해결될 때까지" 중학교 근무를 원하지 않는다고 말했다. 해당 교육구는 WIRED의 논평 요청에 응답하지 않았다.

그는 지역 고등학교에서 대리 근무를 맡았지만 문제는 그를 따라왔다. 한 수업 시간 중, 그는 같은 AI 생성 이미지가 에어드랍(AirDrop)으로 전송되는 것을 받았고, 곧 바로 회수되었다. 그 이미지가 수업 내에서 유통되고 있다는 것을 알 수 있었다. 교장은 학교가 '즉각적인 조사'를 시작하고 있다고 말하며 데산티아고와의 면담이나 학교 상담사 이용을 제안했지만, 데산티아고는 올해 이 교육구로 돌아가지 않을 것 같다고 말한다. 적어도 지금은 그렇게 하는 것이 안전하지 않다고 느낀다고 그는 WIRED에 말했다.

"모든 것을 의심하게 만든다"

동의하지 않는 성인의 가짜 누드 이미지가 유포될 경우, 이는 사이버 스토킹에 해당할 수 있으며, 비동의 성적 이미지 게시를 범죄로 규정하는 연방법 'Take It Down Act'(테이크 잇 다운 법)을 위반할 수 있다.

원문 보기
원문 보기 (영어)
Comment Loader Save Story Save this story Comment Loader Save Story Save this story For Luis DeSantiago, a substitute teacher in Los Banos, California, it started with a few text messages. On March 12, as he was working his other part-time job at a coffee shop, his phone lit up. A fellow barista said there was an AI photo circulating on social media. Her cousin wanted to know if it was DeSantiago. At first, DeSantiago wasn’t worried—he figured that inappropriate photos violated TikTok’s community guidelines and were usually quickly removed. “When I finally saw the image, that’s when it felt like a punch to the gut,” DeSantiago tells WIRED. Parts of it resembled a mirror selfie that he remembered taking at a job in college. But instead of his work uniform, he was wearing lingerie. And instead of standing normally, as he had been in the original photo, he was posing suggestively. The texts kicked off a monthslong saga that crossed two schools and involved students and administrators. And DeSantiago is far from the only educator who’s been victimized this way. WIRED spoke with three other teachers who allege that their students made sexualized photos and videos of them, likely using AI. They say the incidents have been profoundly distressing and disappointing—and in some cases, it’s led them to reassess how they view their work and their relationships with their students. “This Could Include Other Students” After he received the texts, DeSantiago began to piece together who might have been behind the image. One account that shared it used their real name in their TikTok profile—it was a student DeSantiago knew from the local middle school, where he regularly filled in. DeSantiago suspected the involvement of two other students based on people who followed and interacted with that account, and his knowledge of student friend groups. He was scheduled to work at that very middle school the next day. DeSantiago reported the incident to the assistant principal, who assured him that the school would "take care of it.” Later that evening, he decided to look for the image by searching his name on TikTok. He discovered that a different student had created an account impersonating him—it used his name, slightly mispelled, and had the same photo as their profile picture. Over the next few weeks, DeSantiago says, he went back and forth with school administrators, the parents of the students, and even the police. “If these students aren't afraid of generating and posting these images of me, I am concerned of what they may be circulating privately through text messages and group chats,” DeSantiago wrote to the principal. “Keep in mind this could include other students as well.” One student, who admitted to creating the image, was disciplined with in-school suspension. DeSantiago was later told that the image was also shared on Snapchat. But DeSantiago says the image continued to proliferate, and other students began posting it. He says he also received several crude, sexual comments on his TikTok account. DeSantiago told administrators he didn’t want to take assignments at the middle school “until this matter is fully addressed.” The district did not respond to WIRED’s request for comment. He took a substitute assignment at the local high school, but the problems followed him. During one class, he received an AirDropped photo of the same AI-generated image before it was quickly revoked. He could tell it was circulating among the class. The principal told him the school was launching an “immediate investigation” and offered to meet with DeSantiago or make the school counselor available, but DeSantiago says he probably won’t return to the district this year. He tells WIRED that, at least right now, he doesn’t feel safe doing so. “It Makes You Question Everything” When fake nude images of nonconsenting adults are disseminated, it could constitute cyberstalking and violate the federal Take It Down Act , which makes it a crime to publish nonconsensual “intimate visual depictions of individuals, both authentic and computer-generated,” on the internet. In most of the accounts shared with WIRED, the teachers didn't directly report the content to the platforms on which it appeared. Snap spokesperson Monique Bellamy shared the company's policies on harassment and sexual content , which prohibit sharing AI-generated or manipulated content to humiliate or sexually harass others. She told WIRED that the company has "established processes" for reporting such content, and that it continues to "develop features and tools designed to support the safety, privacy, and well-being of our community." TikTok did not provide comment before publication. Donna Ballman, a retired employment advocacy lawyer, tells WIRED that if a student creates and shares sexual images of a teacher, it constitutes a “clear case” of sexual harassment in the workplace. Ballman says that under federal antidiscrimination laws, employers have an obligation to create a workplace free of sexual harassment. In a school, this could mean disciplining the student or moving them to a different classroom. She adds, however, that there's “no blueprint” for how schools need to respond to these incidents. Administrators don’t even have to tell a victimized teacher how a student may have been disciplined. “They just have to make [and] create a safe environment for the victim,” Ballman says. One middle school teacher in her mid-twenties tells WIRED that she’s seen so much AI-generated content used to harass teachers that she’s bracing for the next wave. (She requested anonymity because she did not have permission from her school district to speak about her experience.) Last year, she says, students found public photos of a teacher in her sixties, manipulated them to make it appear that she was committing sex acts, and shared them with each other in a Snapchat group chat. A few months ago, the teacher says, she learned from a student about a TikTok account impersonating her. It looked AI-generated—her face had the characteristic smoothness of AI-generated content, her mouth was moving unnaturally while speaking, and the sound of her voice didn’t match her own. Some of the videos showed her doing things like skydiving or backflips. But others depicted her talking about sexual acts she performed and enjoyed. She says she told her principal about the account and that they assured her that they would “see what we can do.” Throughout the day, students continued to warn her about the account. She says that some told her that they knew who made the account—one of their classmates. She says the student was known to have behavioral problems. “I know a lot of the kids that reported it to me genuinely believed that was my account, and that I was saying those things,” she tells WIRED. “I was like, ‘Oh my god, you guys—that is not me.’” She asked them to report it to TikTok, and she says the account was quickly taken down. Still unnerved by the incident, she mentioned the TikTok account during the next monthly teacher meeting. A colleague said he had seen a seemingly AI-generated fake video of himself getting arrested, which he believed was generated by the same student. As days and weeks passed, she says, she became the target of escalating, in-person harassment from this same student, who began “cursing me out, calling me names, telling me to fuck off, calling me a bitch.” She raised it at the next faculty meeting, and the assistant principal said that the same student was calling and texting them in the middle of the night and posting apparent AI-generated photos of them to TikTok. She says the principal removed the student from her class and tried to limit his access to her. But she says that the harassment continued until the student ultimately left the district, and TikTok removed the account. She says that these behavioral issues aren’t exclusive to this particular student. “It's kind of unsurprising that we