Govt won't oppose Labour bill to tackle AI abuse imagery

Labour’s proposal would update Ireland’s Harassment, Harmful Communications and Related Offences Act. Its central change is to state clearly that sexual abuse images generated or altered using AI fall within the law. This matters because existing rules may not clearly cover synthetic images made without a victim’s participation.

The Bill would make it an offence to generate such images without consent when intending to publish or distribute them. It would also criminalise encouraging or knowingly facilitating their generation for that purpose. The proposal was updated after the Vector controversy, which involved AI-assisted “nudification” of images.

It would extend accountability beyond individual creators. Alan Kelly said companies providing these tools, including major platforms such as X, Meta, TikTok and Google, should answer for their conduct. Government will not oppose the Bill, saying its goals fit a zero-tolerance approach to domestic sexual and gender-based violence.

AI-generated sexual abuse imagery is sexualised material produced by an artificial intelligence tool, often using an existing photograph as a starting point. AI-altered imagery changes a real image, while fully generated imagery may create a fictional-looking scene. The abuse comes from making or using sexualised content without the person’s consent.

“Digital nudification” refers to using AI to alter a clothed person’s image so it appears to show nudity. The source links this practice to the Vector tool and says sexually manipulated images, including images of children, caused public outcry. The proposed Bill would explicitly cover images someone generates or alters with AI.

The article does not provide a technical explanation of how these systems work. It does establish the legal concern: creating such material without consent for publication or distribution, or knowingly helping others generate it, should attract criminal accountability. The proposal therefore focuses on the harmful act and its intended circulation.

The proposal specifically names four major companies or platforms: X, Meta, TikTok and Google. Alan Kelly described them as major publishers and argued that they must answer for conduct connected with harmful AI-generated sexual imagery. The article does not give a separate legal classification for each service.

Their services differ, but the proposal concerns the roles they may play in online publishing, social-media distribution and access to AI tools. X hosts the Vector tool, while Meta and TikTok operate major social platforms. Google is also named among the companies that Kelly says should face responsibility under the law.

The key point is not that all four are accused of the same conduct. Rather, Kelly argues that companies should not claim to take safety seriously while escaping responsibility for facilitating abuse. The Government’s decision not to oppose the Bill means this issue can continue through the legislative process.

The Bill looks beyond the person who presses a button because AI tools and online platforms can make abusive imagery easier to create and spread. Kelly argues that companies providing these tools should share responsibility when their systems help enable non-consensual sexual abuse. This matters because victims may be harmed even when they never participated in creating the image.

Its proposed mechanism has two parts. It would criminalise encouraging or knowingly facilitating image generation for publication or distribution. It would also allow responsibility to reach executives of companies that habitually permit harmful material on their platforms. Kelly specifically argued that ordinary fines do not sufficiently affect companies that build them into their baseline costs.

The article does not say that every platform would automatically be criminally liable. It describes a proposal focused on knowing facilitation and repeated failure to prevent harmful material. Government will not oppose it, while several other Bills addressing the same problem are also before the Dáil.

The article says Labour’s Bill would close gaps in Irish law, but it does not list every existing gap or quote the current Act’s precise limits. Its main proposed clarification is that sexual abuse images generated or altered with AI would be explicitly covered by the Harassment, Harmful Communications and Related Offences Act.

The proposal would address several activities that the Bill’s supporters say need clearer treatment. These include generating images without consent for publication or distribution, encouraging or knowingly facilitating that generation, and the AI-assisted “nudification” of images. It would also strengthen legal accountability for online platforms and their executives in specified circumstances.

The article presents this as part of a wider set of statutory efforts. Minister Colm Brophy outlined existing measures and noted that several Bills are before the House. Alan Kelly said any one of them, or a combination, could be used, provided legislation is enacted quickly.

If enacted, the Bill would make certain conduct criminal. A person who generates an AI sexual abuse image without consent for publication or distribution could commit an offence. So could someone who encourages or knowingly facilitates that generation for the same purpose. The article does not specify the exact penalties for individuals.

The proposal also seeks consequences for platforms and the companies behind them. Kelly said major publishers would have to answer for their conduct, and that executives of companies habitually allowing harmful material could be held criminally liable. He argued that financial fines alone may not change company behaviour because firms can treat them as routine business costs.

These outcomes remain proposed, not current law. Minister Colm Brophy confirmed that the Government would not oppose the Bill after Cabinet approval. The article also says several Bills address the problem, so the final legal consequences could depend on which proposal, or combination of proposals, is enacted.

Consent is the central protection in this proposal. A person’s photograph or likeness does not amount to permission for someone to create, alter or distribute sexual material involving them. Privacy and personal autonomy are harmed when an image is made or circulated without the person’s agreement, especially when the material is presented as real or targets children.

Technology separates the victim from the abusive act. Someone can generate or alter an image without the victim participating. A platform may then provide the tool, host the material or distribute it. The proposed response is to criminalise non-consensual generation for publication or distribution, plus encouraging or knowingly facilitating that conduct.

Intermediary liability concerns when companies should be responsible for users’ conduct. The article supports responsibility for companies that provide tools and for executives where firms habitually allow harmful material, but it does not state the final legal test or penalties. Those details would depend on the enacted legislation.