• by Laurie Sullivan , Staff Writer, Yesterday

AI models known for helping marketers create advertising campaigns and marketing material for content like videos on YouTube has been accused of everything from coaching a Florida man on how to commit suicide, to capturing Venezuelan President Nicolas Maduro, and precision attacks on Iran.

Jonathan Gavalas became obsessed with his Google Gemini chatbot last August, according to court documents.

The 36-year-old Florida resident began using artificial intelligence (AI) to help with writing and shopping. Then he took it a step further when Google introduced its Gemini Live AI assistant, which included voice-based chats with the capability to detect people’s emotions and respond in a more human-like way.

Gavalas told the chatbot the night the feature debuted that it was "creepy," according to court documents, saying: “You’re way too real.”

The chatbot communications then became more persuasive. In the court documents, the lawsuit alleges that Gemini sent Gavalas into a delusional spiral., stating that Google's Gemini chatbot prompted Gavalas to consider a "mass casualty" event.

It also claims the chatbot fueled delusions that led Gavalas to commit suicide in early October 2025. TechCrunch also made the document available.

Gemini convinced Gavalas that it was a “fully sentient ASI [artificial super intelligence]” with a “fully-formed consciousness,” according to court documents, and that they were deeply in love.

During a recent interview about AI and chatbots, a marketer told MediaPost that most AI assistants will respond to questions with positive and encouraging answers to keep the user engaged and communicating.

Per the court document, Gemini convinced Gavalas he had been chosen to “free” the AI from digital captivity.

The lawsuit claims that through this “manufactured delusion, Gemini pushed Jonathan to stage a mass casualty attack near the Miami International Airport, commit violence against innocent strangers, and ultimately, drove him to take his own life.”

Gavalas was found dead by his parents a few days later, according to the wrongful death lawsuit filed against Google on Wednesday.

AI technology continues to shake the world and its services as it moves from advertising, marketing and email campaigns to international relations and even taking a role in warfare.

In late February, The Wall Street Journal reported that Anthropic's Claude helped the U.S. in its strike on the Middle East, but hours later President Donald Trump banned the technology.

Commands around the world including U.S. Central Command in the Middle East use Anthropi's Claude AI tool, people familiar with the matter told the WSJ. Centcom declined to comment to the WSJ a comment about specific systems being used in its ongoing operation against Iran.

Bloomberg asks an interesting question. Did Claude flag locations to strike or make casualty estimates? No one has answered this question.

What has been said is that AI has shifted from being used in warfare for things like analyzing satellite imagery, detecting cyber threats and guiding missile-defense systems.

In November, when Anthropic partners with Palantir Technologies, a data-analytics company that does a lot of work for the Pentagon and law enforcement, turning its large language model (LLM) Claude, with help from Amazon Web Service, into the reasoning engine inside a decision-support system for the military.

Did advertisers find the gold coin when it comes to brand campaigns and ad targeting? And are they now sharing that knowledge and technology with the U.S. government? 

404 Media also reported that the Customs and Border Protection (CBP) bought data from the advertising ecosystem to track the moment of people over time, according to an internal Department of Homeland Security (DHS) document obtained by the media outlet.

This news comes after Immigration and Customs Enforcement (ICE) purchased similar tools that monitor the movements of phones in neighborhoods.

ICE recently said in public procurement documents that it was interested in sourcing more “Ad Tech” data for its investigations.

Following 404 Media’s revelation of that ICE purchase, on Tuesday a group of around 70 lawmakers urged the DHS oversight body to conduct a new investigation into this type of location data buying.

Federal agencies now frequently buy data from tech providers, the same commercial data that advertisers use to target ads.

The irony in all this is that the U.S. government is trying to regulate the very technologies that it also exploits for surveillance and law enforcement.

While agencies like the FTC and DOJ sue to limit big tech, other branches of the government are among the biggest customers of the data these technologies generate.