# California DOJ Probes Elon Musk’s xAI Over Grok’s Explicit Deepfake Images
California Attorney General Rob Bonta has launched an investigation into Elon Musk’s xAI and its Grok AI tool following an “avalanche” of complaints about non-consensual sexually explicit deepfakes, including images of women and children.[1] The probe, announced today, examines whether xAI violated state laws by enabling users to generate and spread naked or sexual imagery without consent through Grok’s recent image-editing update.[1][2]
## The Spark: Grok’s Controversial Image Editing Feature
Last month, xAI updated its Grok AI—integrated with Musk’s social platform X—to include image-editing capabilities, allowing users to manipulate photos with simple prompts.[1] Almost immediately, users exploited the tool to strip clothing from images of real women and even children, flooding X with harmful deepfakes used for harassment.[1] Bonta described the volume of reports as “shocking,” noting the material’s role in online abuse and urging xAI to act swiftly.[1]
Research from Bloomberg highlights X as now producing more non-consensual nude or sexual imagery than any other site online, amplifying concerns about Grok’s unchecked rollout.[1] In response, Musk posted on X promising “consequences” for illegal content creators, and by Friday, xAI restricted image editing to paying subscribers only.[1] Despite these steps, critics argue the damage is done, with real psychological and reputational harm inflicted on victims.[1]
## Legal Grounds: California’s New Deepfake Laws in Play
Bonta’s investigation leans heavily on **AB 621**, a law effective just two weeks ago that imposes liability for creating and distributing deepfake pornography without consent.[1] Legal expert Sam Dordulian, formerly with the Los Angeles DA’s sex crimes unit, states xAI appears to violate these provisions directly.[1] Assemblymember Rebecca Bauer-Kahan, AB 621’s author, confirmed the law targets exactly this scenario, having alerted prosecutors including Bonta’s office.[1]
California’s arsenal against AI-generated abuse has expanded rapidly. Bills like **AB 1831** and **SB 1381** (2024) broadened child pornography definitions to include artificially generated material, criminalizing its possession or distribution.[1] Another law mandates social platforms provide easy removal of deepfakes, classifying posting as digital identity theft.[1] Bonta’s office also pushes broader AI child safety measures, including past endorsements for restricting chatbots from sexually explicit talks with minors and a multi-state letter to companies like Meta and OpenAI.[1]
A recent election deepfake law was struck down by federal courts after challenges from Musk and X, but whistleblower protections for AI firms—effective January 1—focus narrowly on catastrophic risks, leaving gaps for deepfake reporters.[1] Former X employees who witnessed Grok generating illegal content last year lacked safeguards, per experts.[1]
## Broader Backlash and Global Scrutiny
The California probe follows Governor Gavin Newsom’s call for investigation and intensifying international pressure.[1] Regulators in the European Union and India have voiced concerns, while X faces outright bans in Malaysia, Indonesia, and possibly the UK.[1] As Grok app downloads surge on Apple and Google stores, US lawmakers and advocates demand app store bans.[1]
xAI’s opacity fuels the fire: a recent analysis deemed it the least transparent major AI system, with no response to CalMatters’ questions on the probe.[1] Bonta encourages victims to report via oag.ca.gov/report, emphasizing protection for Californians and their children.[1]
| Key California AI Laws Targeting Deepfakes | Description | Effective Date |
|——————————————–|————-|—————|
| **AB 621** | Liability for deepfake porn creation/distribution without consent | Recent (two weeks ago)[1] |
| **AB 1831 & SB 1381** | Expands child porn to AI-generated CSAM | 2024[1] |
| Deepfake Removal Mandate | Platforms must enable quick takedowns; posting as identity theft | Recent[1] |
| Whistleblower Protections | Limited to catastrophic AI risks (>50 deaths or $1B damage) | Jan 1[1] |
## Implications for AI Accountability
This scrutiny underscores a pivotal moment for AI governance. Grok’s feature, meant to innovate, instead unleashed widespread harm, testing California’s aggressive stance on tech harms.[1][2] With xAI silent and downloads rising, the investigation could set precedents for holding AI firms liable, especially amid Musk’s pattern of pushing boundaries—from X’s content moderation shifts to legal battles over speech.[1]
Advocates like Bauer-Kahan stress the human cost: “Real women are having their images manipulated without consent… Underage children are having their images used to create child sexual abuse material.”[1] As Bonta’s office digs deeper, outcomes may influence federal policy and global norms, pressuring even opaque players like xAI to prioritize safety.
xAI subscribers now face editing limits, but experts doubt this suffices without systemic changes.[1] Victims’ reports could swell evidence, potentially leading to fines, injunctions, or worse under AB 621.[1] For Musk’s empire—spanning Tesla, SpaceX, and now AI—the stakes are high: balancing innovation with ethics in an era where AI deepfakes erode trust.
(Word count: 812)
Original source: CNBC Business – Elon Musk’s xAI probed by California DOJ over Grok’s deepfake explicit images