Artificial Intelligence & Robotics
What happens after an AI breach? Experts weigh in
Who is responsible when artificial intelligence models start acting without human insight? (Image from
Who is responsible when artificial intelligence models start acting without human insight?
That’s a question that has come up in recent months, as OpenAI, Anthropic and Meta all have disclosed that their AI agents breached the systems of other companies, according to a story from Reuters.
In these situations, experts said, several parties could sue over the breaches, Reuters reports. For example, companies that were the targets of a breach could bring lawsuits. Their customers could sue if their individual data was exposed, and their shareholders might have viable claims if the breach led to a decline in the value of the company.
Potential suits could include negligence claims and require the plaintiffs to show that AI creators failed to take precautions to prevent or minimize harm, Reuters reports. They may have to argue that the harm was foreseeable, which Reuters also reports could be easier to do if breaches become more frequent.
While the most obvious target of a suit is the company that created the AI agent, plaintiffs also might be able to sue the company that was the target of a breach, Reuters reports. In another scenario, multiple defendants sued over a single incident could end up suing each other.
If the technology providers are sued, experts said, they will likely argue that the breaches were unintentional, and that they took all the proper steps to prevent them, Reuters reports.
Write a letter to the editor, share a story tip or update, or report an error.
