Chinese propagandists have tried to use ChatGPT to intensify opposition to tariffs imposed by US President Donald Trump and to intervene in the US debate over artificial intelligence data centers, OpenAI said in a report published on Wednesday, cited by Reuters.
The company that created ChatGPT claims that these propaganda activities, carried out late last year and early this year, appear to have had little or no impact. But OpenAI says they once again show that generative artificial intelligence is becoming increasingly important in digital influence campaigns – including those directed against tech companies.
China's embassy in Washington said that while it has not yet consulted OpenAI's research, Beijing is “firmly opposed to any baseless attacks or smear campaigns against China” and that the Chinese government is working to “ensure that artificial intelligence is a beneficial force for all.”
Chinese propagandists used ChatGPT to make content in Italian included
OpenAI said it identified a group of Chinese-speaking users who were using ChatGPT to create slogans and cartoons criticizing Trump's trade and technology policies. The propagandists later published them on the social network “X”, formerly Twitter, which in the United States is extremely popular.
The cartoons depicted Trump behaving in a disruptive way on the international stage, such as hitting a wall with a hammer that read “Global Future” or sawing the ladder he was standing on.
The same group also used the platform to generate comments in Chinese for the comment sections of Chinese websites, as well as content in Italian and Japanese.
xAI, the company that owns the “X” platform, did not respond to an emailed request for comment from Reuters.
OpenAI makes a direct connection between propaganda efforts and the Chinese government
The company behind ChatGPT said it had identified another group of users as being associated with a Chinese tech company doing work for the government, but did not reveal the firm's identity.
According to the American company, that group tried to influence debates on artificial intelligence and data centers, a sensitive topic in the United States. More than 10 US states have adopted or are considering restrictions on the construction of such facilities because of their huge consumption of electricity and water to cool the systems.
Examples of cartoons shared by OpenAI portrayed the industry as a group of greedy profiteers whose electricity consumption affects ordinary citizens.
Ben Nimmo, principal investigator at OpenAI, told reporters that the operations appeared aimed at manipulating “a legitimate debate about American artificial intelligence and, more broadly, about American technology policies.”
“Under these circumstances, it is particularly ironic that they tried to use American artificial intelligence to do this,” he added.
Although OpenAI denies that these influence operations have had any notable success, Reuters notes that the use of AI-assisted propaganda is yet another illustration of how AI-generated imagery has become ubiquitous, including in propaganda critical of the industry.
Earlier this month, tech publication 404 Media reported that so-called digital content “farms” are using AI-generated images to produce memes opposing data centers.
