This article is more than 1 year old
China follows through on plan to ban deepfake tech
Synthetic media undermines national security, says internet regulator
China's new rules banning the creation of AI deepfakes used to spread fake news and impersonate people without consent will take effect on Tuesday.
Government policies enforced by the Cyberspace Administration of China (CAC) will require "deep synthesis service providers" to ensure their AI algorithms are not being misused for illegal activities, like committing fraud, scams, or distributing false and harmful information. They must also clearly label synthetic content by adding a watermark.
The CAC said synthetic media threatens China's social stability and national security. The internet regulator warned against using AI to "produce, copy, publish, and disseminate illegal and harmful information" or impersonating another person's identity to commit fraud, according to a translation of a statement issued last month.
The new rules outlined in the "Regulations on the Administration of Deep Synthesis of Internet Information Services" come into effect on 10 January 2023. Online identities should be authenticated by platforms, and users should be able to report suspicious fake content. Meanwhile, companies must protect user data to prevent sensitive information from being used to craft malicious deepfakes.
"Services that provide functions such as intelligent dialogue, synthesized human voice, human face generation, and immersive realistic scenes that generate or significantly change information content, shall be marked prominently to avoid public confusion or misidentification. It is required that no organization or individual shall use technical means to delete, tamper with, or conceal relevant marks," the CAC added.
- China bans deepfakes created without permission or for evil
- Real-time deepfakes can be beaten by a sideways glance
- Deepfakes, quantum computing cracking codes, ransomware... Find out what's really freaking out Uncle Sam
- Do you expect me to talk? Yes, Mr Bond, I expect you to reply: 10k Brits targeted on LinkedIn by Chinese, Russian spies
Deepfake technology has dramatically improved over the years, leading to commercial applications featured in adverts and films. Although it offers new creative techniques, many question whether it's fair to imitate someone's likeness by training a model on their photographs and speech without permission. Some thought it was wrong when the voice of Anthony Bourdain was cloned in a documentary about him after he had passed away, for example.
The EU also has rules in place restricting deepfakes online, including fining companies who do not have policies in place to detect and remove disinformation and bot accounts. Some states in the US have banned using deepfakes to influence elections or make revenge porn. ®