BRUSSELS (Reuters) -Apple, Meta Platforms and other tech giants may win a reprieve from the EU's landmark artificial intelligence rules as regulators consider easing sections of the legislation as part of a drive to simplify a slew of regulations adopted in the last two years.
The move by the European Commission came amid intense lobbying by big tech companies and criticism from the U.S. administration against the AI Act adopted last year, which applies risk-based rules to artificial intelligence.
EU tech chief Henna Virkkunen will present the so-called Digital Omnibus on November 19 according to a Commission agenda. The document could still be changed before then.
"The Commission is proposing targeted simplification measures aimed at ensuring timely, smooth and proportionate implementation," the draft Digital Omnibus document seen by Reuters said.
The changes include exempting companies from registering their AI systems in an EU database for high-risk systems if these are only used for narrow or procedural tasks, and the introduction of a one-year grace period where authorities can only levy penalties from August 2, 2027.
A requirement for AI system providers to mark their output as AI-generated content to address concerns such as deepfakes and misinformation will be subject to a transitional grace period, the document said.
The EU executive has in recent weeks watered down landmark environmental rules after blowback from companies and the U.S. government.
(Reporting by Foo Yun CheeEditing by Gareth Jones)
