The Ugly Side of A.I.: How Manipulated Images and Audio Are Being Used For Online Harassment, Hate Campaigns, and Misinformation
In October, the Louisiana parole board convened to discuss the potential release of a convicted murderer. It called on a mental health expert to discuss the inmate. However, while the parole board was meeting, a collection of online trolls took screenshots of the doctor from an online feed of her testimony and, using A.I. technology, edited the images to make her appear naked. These manipulated images were then shared on 4chan, a notorious anonymous message board known for fostering harassment and spreading hateful content.
According to Daniel Siegel, a graduate student at Columbia University, such incidents involving new A.I.-powered tools like audio editors and image generators being used for racist and offensive content on 4chan are not isolated. Mr. Siegel chronicled the activity on the site for several months and found that nefarious users could employ sophisticated A.I. tools to supercharge online harassment and hate campaigns.
Fringe message boards like 4chan often give early warning signs for how new technology could be used. These platforms are filled with young people who are quick to adopt new technologies like A.I. to project extreme ideas, says Callum Hood, the head of research at the Center for Countering Digital Hate.
Some problems resulting from A.I. tools that experts discovered on 4chan include the creation of fake pornography through image generators, and cloning of voices using A.I. technology. While the manipulated images and audio have not spread far beyond 4chan, the Louisiana parole board has opened an investigation in response to these findings.
However, experts have pointed out that there are no federal laws banning the creation of fake images of people. While Illinois and other states have passed laws to ban the distribution or creation of A.I.-generated pornography without consent, much still needs to be done to regulate and monitor the spread of such content.
Further, the misuse of A.I. technology by individuals on 4chan is not restricted to manipulated images and audio. For example, when an A.I. company released a tool that could create a convincing digital replica of someone’s voice, 4chan users circulated clips of fake celebrity voices.
Considering the broader implications of A.I. misuse, major social media companies have since required labels on some A.I. content. President Biden also issued an executive order directing the need for labelling such content, and the development of standards for watermarking and authenticating A.I. content.
Meta, formerly known as Facebook, has adopted an open source strategy to release its software code to researchers. However, on 4chan, the leaked code was used for different ends, including creating new chatbots capable of producing antisemitic ideas. This highlights the risks associated with free-to-use and open-source A.I. tools being tweaked by technologically savvy users.
In conclusion, while A.I. has the potential to revolutionize industries and improve lives, the misuse of A.I. tools poses serious threats. It is critical for regulators and technology companies to take proactive measures to monitor, regulate, and combat the malicious use of A.I. technology to spread hatred, harassment, and misinformation.