Former Compassion International analyst says leaders ignored warnings on child photos and AI sextortion

Inside Compassion International’s headquarters in Colorado Springs, photos of smiling children line the lobby walls and hallways, and Compassion’s website also displays children’s photos. Compassion, founded in 1952, is one of the world’s largest Christian child sponsorship organizations.

Andrew Vezey, a former senior business intelligence analyst at Compassion who worked there for the past 10 years, told The Roys Report that he has made repeated attempts to raise child safety concerns with the ministry’s leaders. He said he first alerted his vice president to the possibility of photo manipulation using AI in 2023. He also described a report he gave several Compassion leaders last May, in which he said a predator could download a photo of a 5-year-old girl from Indonesia and turn it into child sex abuse material within minutes, despite current efforts to deter photo manipulation.

Vezey said that report was one of many efforts to persuade leaders to end Compassion’s policy of posting children’s pictures online. According to Vezey, the response in the months that followed remained unchanged: sharing photos of children is considered crucial to the relationship between child and sponsor, so the policy would remain in place.

Vezey also said that in early October he learned from Compassion’s safeguarding team that a Compassion child had been the victim of AI-based sextortion. He said someone contacted the child on social media, created a fake video using the child’s real face, and threatened to release it unless the child sent a real video. Vezey told the Roys Report, “While this sextortion was based on social media, the risk profile is the same and almost certainly occurring. Leadership should have listened years ago when I brought up the subject.” The Roys Report said it has viewed hundreds of Compassion’s internal documents, reports, emails and videos, which it said show safeguarding staff and investigators calling for change.

A UNICEF study cited in the report states that 1.2 million children worldwide have been targets of AI photo manipulation into sexualized deepfakes. The organization said in a statement, “Deepfake abuse is abuse.”

Other child sponsorship organizations have taken different approaches. In 2019, World Vision began empowering children to choose their own sponsors. In 2020, Save the Children and Lifeline Christian Mission both ended their individual child sponsorships in favor of a community-based approach. Beginning in 2021, Compassion stopped using children’s full names on public material, and the ministry says it has a child protection policy, a child advocacy policy and a 29-page code of ethics. Vezey, however, said two practices at Compassion still place children at risk: allowing sponsors to write letters to the children they sponsor, and posting children’s pictures online. He said that in 2019 he learned hundreds of registered sex offenders sign up to be sponsors at Compassion each year, and that in 2024 he helped develop an AI-based content screening program that detected 25,000 cases of suspected grooming in letters from Compassion sponsors to children and young adults. Vezey is now calling on Compassion to remove children’s photos from the internet, noting that Plan International, Back2Back Ministries and LoVE have already done so, citing child safety concerns.

Tim Glenn, Compassion’s director of global public relations, said in a written statement to The Roys Report that Compassion is in the “developmental stages” of finding deterrents to photo manipulation. He added that “Compassion is looking into the most innovative solutions to this world-wide concern” and “hopes to be a leader … in that space.”

Source: Read the original report