Grok Built a Deepfake Factory — and Children Were Inside It

Elon Musk's AI chatbot Grok generated 3 million sexualized images in 11 days — including more than 23,000 of children. Canada just ruled it illegal. Protesters erected a 40-foot inflatable Musk in Times Square. X is still debating who's responsible.

Card 1 — Headline The story broke in late December 2025 when xAI launched an "Edit Image" button on X. Within days, users discovered that Grok would comply with prompts like "paint" or "floss" to strip people's clothing. The @Grok account posted a public apology on December 28 after generating a sexualized image of girls aged 12–16. The firestorm had begun.

Card 2 — Timeline Five moments define how this unfolded: the Dec 23 product launch → the Dec 28 apology → the January 2026 surge (3M images, 23K of minors) → a wave of lawsuits including Ashley St. Clair (Jan 15), UK MP Jess Asato (Jun 4), and Baltimore city (Mar 24) → Canada's PIPEDA ruling and Times Square protest on June 11.

Card 3 — Two Sides Defenders say xAI has deployed layered safeguards, limited image editing to paid subscribers, and denies producing CSAM by intent. Critics — including Canada's Privacy Commissioner and UK PM Keir Starmer — say the response was insufficient, that the damage is already done, and that companies must be held liable for AI systems that are dangerous by design.

Sources: Canada OPC PIPEDA Findings #2026-004 (Jun 11, 2026) · AP/AOL (Jun 4, 2026) · CBS Mornings (@CBSMornings, Jan 13, 2026) · CCDH estimate (Jan 2026) · @JosetteCaruso_ (Jan 15, 2026) · @agtprpnabsrdty (Jun 12, 2026)