Croatian newspaper publishers are proposing a centralized AI-generated online comment moderation system available to all publishers as a tool to regulate hate speech in user comments on news websites, state agency Hina reported on Thursday.

They sent their proposal to Croatia’s Ministry of Culture and Media, saying that the AI system would help their webmasters sort and remove user-generated comments that deemed insulting or hateful. The publishers are also reportedly looking into tapping EU funds to help develop such a comprehensive solution to hate speech online.

“More information about this will be released to the public after the proposal is discussed at the ministry,” UNI told Hina.

The proposal comes at the time when the Parliament is debating passing a new new media bill, which is likely to include provisions making publishers of online content liable for any hate speech appearing in comments under their articles, effectively forcing publishers to step up moderation of user-generated messages on their websites.

“Adapting to the new definition of publishers’ responsibilities is complex, expensive and demanding – and we need time and money to do that,” publishers told Hina.

The group said they expected the new system to be made available withing the next two years.

The publishers added that in lieu of the proposed system alternatives include automatic delaying of comments before they appear on the website, or more stringent control exercised by human moderators.