Irish publishers and music-industry representatives are putting copyright at the centre of the debate about artificial intelligence. On 14 October 2026, NewsBrands Ireland and the Irish Recorded Music Association are due to hold a conference entitled Copyright and Human Creativity in the Age of AI at the Irish Film Institute in Dublin. The event forms part of Ireland's EU Council presidency programme. It is an upcoming discussion, not an enacted ban on AI tools or a newly announced compensation scheme. For journalists, musicians, independent creators and businesses, the core questions are who can use protected material for AI development, under what conditions and how lawful licensing might work.

What is taking place in Dublin on 14 October?

The conference is scheduled for Wednesday at the Irish Film Institute, Temple Bar, Dublin 2. The organisers are NewsBrands Ireland, which represents national news publishers, and IRMA, which represents the recording industry. The Department of Enterprise, Tourism and Employment has endorsed the event. Its focus is copyright and human creativity in an era when AI tools can generate words, images and music rapidly. Registration is already marked closed on the official conference website. Readers should not confuse this with the separate International AI Summit, which has a different agenda and audience.

Who is due to speak?

The published programme includes European Commission Executive Vice-President Henna Virkkunen, Irish Minister Peter Burke and copyright specialist Professor Eleonora Rosati. The conference is to be hosted by Business Post editor Daniel McConnell. Planned discussions involve media executives, music representatives, policymakers and legal experts. Their presence establishes that the debate has reached a high policy level; it does not mean that they have agreed in advance on a single regulatory solution. Conference remarks should be attributed after they are delivered, not described as completed announcements beforehand.

Why do Irish publishers care about AI training?

Original reporting costs money: newsrooms pay reporters, photographers, editors, legal advisers and technology staff. Some AI models are trained on large collections of text and other material, raising questions about copyright ownership and whether rights holders should authorise use or receive payment. A company that indexes freely accessible web pages is not automatically authorised to reuse every protected work in every way. Equally, not every computer-assisted analysis of text is unlawful. The answer depends on the activity, applicable law, licences and any relevant legal exceptions.

What is the music industry's concern?

Songwriters, performers, labels and music publishers face questions about how recordings and compositions are used in model training, how synthetic songs may resemble existing works and how creators can be credited. Rights can differ between a composition and its recording. A singer's voice, a melody and a sound recording may also raise different legal and ethical issues. It would be inaccurate to claim that every AI-generated melody necessarily infringes copyright. The industry is seeking workable ways to protect identifiable rights while allowing innovation under clear conditions.

Does the EU already have copyright rules?

Yes. European copyright law existed long before today's generative AI systems. Directive 2019/790 addressed copyright in the digital single market and contains provisions concerning text and data mining, including rights reservations in relevant contexts. How those exceptions operate alongside commercial AI training has become controversial. Courts and regulators may have to interpret specific situations. Neither the existence of an exception nor a publisher's ownership automatically answers every question. Irish businesses should distinguish the EU framework, Irish implementation and the precise permissions in the contracts they sign.

How does the AI Act affect training data?

The EU AI Act requires providers of general-purpose AI models to have a policy for complying with Union copyright law and to publish a sufficiently detailed summary of training content using the Commission's framework. These provisions do not grant a universal licence to scrape the internet, nor do they by themselves require payment for every reference to a published article. They create additional transparency and governance obligations for model providers. The AI Act and copyright legislation operate alongside one another and address different aspects of AI development.

Can an independent Irish writer protect work?

A creator can keep dated records of drafts, publication dates, original files and licensing agreements. When selling rights, the contract should clarify whether AI training, text reproduction or automated commercial reuse is included. Publishers may also implement machine-readable rights reservations where relevant, but should not assume a robots.txt rule resolves every copyright dispute. If a work is copied or republished without permission, preserve evidence before contacting the alleged user or obtaining legal advice. Public accusations are no substitute for reviewing the actual rights and facts.

What about schools, researchers and startups?

Research, teaching, journalism and commercial software can involve different circumstances. A nonprofit project may qualify for an exception that a separate commercial reuse does not, but nonprofit status alone does not settle the issue. Irish startups developing AI products should inventory data sources and permissions rather than rely on assertions that material is free because it appeared in search results. Universities should set appropriate governance arrangements for datasets, student work and sensitive information. Clear records reduce uncertainty for creators and developers alike.

Could licensing become an economic opportunity?

Structured licences can create a route for AI developers to access trustworthy material and for creators to be paid under agreed terms. Collective arrangements may reduce transaction costs in some sectors. Yet licensing markets can also favour very large companies unless smaller publishers, freelance writers and independent musicians can participate meaningfully. Practical discussions should cover what is licensed, for which uses, for how long, at what price and with what audit rights. No universal pricing formula has been established by the conference announcement.

What should news readers understand about AI summaries?

An AI-generated overview may be useful for discovering information, but it can omit context, misstate facts or obscure which reporter found the original information. The presence of a citation does not itself prove that the summary is complete or authorised. Readers can support verifiable reporting by opening source articles and checking evidence. News organisations should clearly explain when AI tools assist with editing or production and retain human accountability for factual accuracy. Transparency about the tool and respect for third-party rights are separate duties.

Will the conference produce a new law immediately?

No. A conference can influence negotiations, build consensus and highlight proposals, but legislation requires its own institutional procedures. The Irish government, EU institutions and courts have distinct roles. An idea advanced on 14 October would not automatically change what creators may do the next morning. Watch for official statements, Commission proposals and published legal texts before drawing conclusions about new obligations. The event is part of an ongoing policy debate rather than a final ruling on every AI system.

Practical checklist for Irish creators and organisations

Creators should review publication and distribution contracts, maintain ownership records and understand whether they have authorised machine-learning uses. Publishers can examine rights reservations, licensing systems and how automated crawlers access content. AI developers should document their data sources and the legal basis for usage, and have procedures for rights-holder requests. Businesses purchasing AI services should ask about training-data governance and indemnities. None of these steps guarantees that litigation is impossible, but they can support more accountable decisions.

Frequently asked questions

Is attending the Dublin conference free? Check the official organiser rather than assuming; registration is listed as closed. Is AI banned in Ireland? No. Does every use of published material require a new licence? Legal rights and exceptions depend on the use. Will artists automatically receive royalties after 14 October? No such automatic payment scheme has been enacted by this event. Is the EU AI Act relevant? Yes, especially its copyright-policy and training-content transparency requirements for general-purpose AI providers.

Sources and more information

Official copyright conference programme; Irish government International AI Summit announcement; European Commission guidance on GPAI copyright requirements. Read other stories in Technology.