Artificial intelligence drives hike in computer-generated child abuse images

In 2022, 20 reports of computer-generated child sexual abuse material were traced to Ireland. Picture: PA

There has been a sharp increase in computer-generated child sexual abuse material, driven by freely-available artificial intelligence software, the industry internet watchdog has said.

Hotline.ie said that 37% of all reports classified as child sexual abuse material (CSAM) in 2022 involved computer-generated imagery, compared to 9% of cases in 2021.

While the vast bulk of such material comes from abroad, 20 cases were traced to Ireland, according to Hotline's latest annual report.

It said computer-generated imagery — using everything from digital art tools to AI technology — was typically “more severe” and depicted very young child-like renditions.

Analysts at the internet industry watchdog warn that advancements in AI-generated art were “deeply concerning” and that it can take any image of a child and turn it into sexual abuse material.

They said it can be difficult to tell the difference between AI material and images of real children and that, in some cases, the two are mixed together.

The annual report reveals:

• 31 reports of CSAM were traced to Ireland, the highest on record and compared to seven the year previous;
• Two-thirds (20) of the 31 cases related to computer-generated CSAM;
• For the first time, Hotline.ie produced full-year statistics on intimate abuse imagery, or the non-consensual sharing of intimate images or videos, with almost 700 reports; 
• A “surge” in sexual-extortion scams where people, mostly young men, were deceived into sharing intimate images of themselves and pressured to pay up to avoid circulation online.

Hotline.ie takes referrals from the public regarding ‘illegal content' — the vast majority are child abuse material — and works with An Garda Síochána.

It was also referred 321 suspected financial scams and 108 reports of suspected racism and xenophobia.

The report said there has been a “significant rise” in computer-generated child sexual abuse material, with 5,015 classified cases (37% of the total), compared to 1,329 (9%) in 2021.

Computer-generated CSAM was found to be more severe and depict very young child-like renditions.

The trend is being seen across the world, with the UK Internet Watch Foundation warning last year that AI could create these images “at scale” with the clear potential to “overwhelm” those working to fight CSAM.

Hotline.ie received 31,068 reports in 2022, compared to 29,794 in 2021. This includes 13,799 CSAM, compared to 14,772 in 2021, which was a five-fold jump on 2020.

The ISPCC’s head of policy, Fiona Jennings, commended Hotline.ie.

“Their team of expert analysts has shared their concerns about how AI is being used by bad actors to create child sexual abuse material on a level not seen before, in particular how new CSAM is being created from old videos of children previously being sexually abused," she said.

This ought to concern us all as such activity serves to retraumatise these children whilst those creating it, and at times profiting from it, remain unscathed.

She said it was nearly a year since the Government established Coimisiún na Meán, which has a broad remit to protect people from harmful content.

“Its online safety commissioner is currently consulting on its first binding online safety code, which will compel designated platforms to deal more effectively with such content and with those who generate it," she said.