Google, Meta and TikTok pushed back at a parliamentary hearing on proposed changes to Indonesia’s copyright law, which would require platforms to compensate publishers for using news content and set rules for AI.

The government says the changes would protect local publishers and ensure fair compensation, while the companies warn the rules could limit Indonesians’ access to news and undermine publishers’ earnings.

Meta’s Berni Moestafa argued that mandatory payments could reduce publishers’ readership, advertising and subscription revenue, and said linking to an article is different from reproducing or republishing it; Google’s Agung Pamungkas also warned that charging for news links could restrict access to news.

A July draft would require payments for aggregating, republishing or previewing news links and for using news to train AI, with payments routed through state-supervised collective management organizations and sanctions potentially including loss of local business permits.

The Press Council favors a hybrid royalty system that allows collective payments or direct deals, arguing a collective-only model could create an unreliable or opaque intermediary; Google has also asked to preserve direct agreements.

Google cited existing commercial arrangements, including News Showcase partnerships with more than 30 local publishers; an Indonesian news site reported 30 publisher-rights deals and five AI-training agreements, including deals with Antara and Kompas, with an opt-out for publishers.

The July draft would deny copyright to works generated entirely by AI and protect AI-assisted works only if they meet human-involvement criteria, though the required level of input remains undefined; it also proposes restricting AI imitation of a creator’s distinctive style without explaining how that would be assessed.

TikTok’s Noudhy Valdryno urged lawmakers to distinguish fully AI-generated material from work made by people using AI for tasks such as editing, translation or image enhancement, and said disclosure rules should not apply to assistive uses.

Valdryno also called for safe-harbour protections for user-generated-content platforms, which cannot always verify copyright ownership or licensing.

The government and parliament have agreed that works created entirely by AI would receive no copyright protection, but questions remain over disclosure rules, payment models and whether permit revocation will survive in the final bill.

The debate reflects wider disputes over whether AI companies may use human-created work without permission or compensation, with supporters saying stronger rules could give publishers, journalists and other creators more control.

Lawmakers and the intellectual property directorate agreed to accelerate deliberations, forming a working committee on October 6; government negotiators are preparing 1,117 review items and 736 explanatory notes across four policy clusters.