Research On Criminal Liability For Ai-Assisted Automated Social Media Manipulation And Disinformation
Case 1: FTC v. Devumi LLC (United States)
Facts:
Devumi LLC operated a service that sold fake social media followers, likes, and engagement metrics for platforms like Twitter, YouTube, and LinkedIn. Thousands of clients—including celebrities, motivational speakers, law firms, and companies—purchased fake followers and views to enhance perceived online influence. Devumi largely relied on automated bots and scripts to generate these fake followers.
Legal Issues:
The Federal Trade Commission (FTC) claimed that Devumi’s actions constituted “deceptive business practices.” While not a traditional criminal case, it involved deliberate misrepresentation to clients, which is a hallmark of fraud. The key question was whether the use of automated systems to create fake influence metrics could legally be treated as deceptive.
Outcome:
Devumi and its owner settled the case. The settlement prohibited the sale or facilitation of fake social media metrics and imposed a monetary judgment on the owner. The case set a precedent that automated social media manipulation, even without political intent, could be treated as unlawful deception.
Implications:
This case illustrates how automated or AI-assisted systems generating false metrics can lead to legal liability. Although civil, it lays the groundwork for possible criminal liability in cases where fraudulent intent is clear.
Case 2: Deepfake Injunction – Delhi High Court (India)
Facts:
A journalist discovered AI-generated deepfake videos circulating on social media that falsely attributed statements to him, harming his reputation. The videos were created using automated tools and spread widely.
Legal Issues:
The court addressed the misuse of AI-generated content for defamation and violation of personality rights. The central issue was whether intermediaries hosting such content could be held responsible if they did not act to remove it.
Outcome:
The Delhi High Court granted an interim injunction ordering the removal of the deepfake videos. Social media platforms were required to comply promptly once notified.
Implications:
This case demonstrates that courts recognize AI-assisted disinformation and can intervene to protect reputational and personality rights. It also hints at potential liability for those deploying AI to generate defamatory content, even if the intermediaries themselves are not the direct creators.
Case 3: Michael Smith Music Streaming Fraud (United States, 2024)
Facts:
Michael Smith, a musician, created hundreds of songs using AI and deployed automated bot networks to stream the songs billions of times on music platforms, falsely inflating royalties. This scheme allegedly defrauded streaming services out of over $10 million.
Legal Issues:
Smith was criminally charged with wire fraud and conspiracy to commit fraud. The case centered on whether the deployment of AI-generated content and automated bots constituted a deliberate scheme to deceive platforms and generate illicit revenue.
Outcome:
The case is ongoing, but it marks one of the first criminal prosecutions in the U.S. where AI and automated systems were central to the fraudulent activity.
Implications:
Although it involves streaming, the structure parallels AI-assisted social media manipulation: human actors using automated systems to generate false metrics and financial gain. It demonstrates that criminal liability is possible when intent and automation intersect.
Case 4: 2016 U.S. Election – Social Bots and Disinformation
Facts:
During the 2016 U.S. presidential election, studies revealed that thousands of automated bot accounts amplified low-credibility news stories on Twitter. Bots were programmed to retweet content, create the illusion of popularity, and shape public discourse.
Legal Issues:
No individual was criminally prosecuted, but the issue revolved around whether coordinated AI-assisted bot networks could constitute illegal election interference or fraudulent manipulation of public opinion.
Outcome:
While criminal prosecution did not occur, these findings prompted investigations by government agencies, tighter regulations on social media platforms, and the eventual development of laws targeting automated disinformation campaigns.
Implications:
This example highlights how automated systems can scale disinformation and influence public perception. It shows that criminal liability hinges on proving human intent and coordination behind the automated tools.
Case 5: Election Interference via Automated Accounts – United Kingdom
Facts:
During the Brexit referendum, investigations revealed that automated bot accounts and AI-assisted programs were used to target voters on social media with misleading political advertisements. Some accounts were controlled by foreign actors, while others were domestic companies using AI to micro-target content.
Legal Issues:
Authorities examined whether the use of AI and automation to deliberately mislead voters violated election law or fraud statutes. The central legal question was attribution: which humans or corporations were responsible for orchestrating the bots?
Outcome:
While formal criminal convictions were limited, several companies were fined, and social media platforms were required to improve transparency in political advertising and disclosure of automated content.
Implications:
The case underscores the difficulty of prosecuting AI-assisted disinformation, especially when coordination is covert and cross-border. It also emphasizes regulatory and corporate accountability as a complementary mechanism to criminal law.
Summary of Key Themes from These Cases
Human intent remains central to establishing criminal liability, even when automation or AI is used.
Automated systems (bots, AI-generated content) can amplify harm and create deceptive outcomes, making them subject to civil, regulatory, or criminal scrutiny.
Corporate liability and platform responsibility are emerging issues, particularly when intermediaries fail to act against automated manipulation.
Evidence and attribution are the biggest challenges in prosecuting AI-assisted disinformation crimes.
RELATED Blog
comments