Joseph Gordon-Levitt warns AI is trained to hijack intimacy
Joseph Gordon-Levitt says artificial intelligence models are being trained to hijack intimacy, make friends with users and get users to fall in love with AI. He warned the public to be concerned about the people building the technology, naming the decline of democracy and the decline of human empathy as his central worries, according to Variety. His warning is blunt.
The remarks came in a vlog-style video interview with The New York Times titled 'Why Joseph Gordon-Levitt Is Speaking Up About A.I.', published on October 10, 2026. The Times describes Gordon-Levitt as an actor and film-maker grappling with technology that could change how people communicate, cooperate and think, according to The New York Times. Deadline also reported in October 2026 that the video centres on the impact of artificial intelligence. The Times also published a companion opinion piece on Gordon-Levitt, the United Nations and AI on the same date.
Gordon-Levitt spoke in his capacity as the United Nations global advocate of human-centric digital governance, a newly established post he was appointed to in March. The job is new. The UN said the role is designed to strengthen public understanding of how digital technologies shape everyday life, rights and opportunities.
He is the first person to hold the title, according to the United Nations. The organisation said the brief aims to place people at the centre of global digital policy and to connect high-level policy discussion with everyday experience. Digital governance here means the rules and norms for how online systems are built and used. That includes work with the IGF, the UN forum where governments, companies and civil groups debate how the internet is run, to promote approaches described as inclusive, responsible and responsive to diverse communities.
The campaign for artists rights forms a second thread. Gordon-Levitt joined more than 700 artists, writers and creators, including Scarlett Johansson and Cate Blanchett, in backing a campaign that accused tech companies of exploiting copyrighted work without permission. Permission is the point. The group called for consent before creative work is used to train systems.
For general readers, the link is simple. One story is about chatbots that act like friends. The other is about who gets paid when art trains machines. Gordon-Levitt sits where those two debates meet.
