Google is investing his money where his AI is. The company has announced a new bug bounty program that is fully focused on artificial intelligence, inviting security researchers and moral hackers to find flaws in their AI systems. Depending on how serious the issue is, Google is offering awards that are high as $ 30,000.

This new initiative expands Google’s long -standing vulnerability reward program, but focuses on the growing world of AI, where security threats look very different from traditional bugs. The company says that it wants to help experts to identify “wicked actions”, when AI behaves in unexpected or dangerous ways, such as leaking personal data, executing incorrect orders, or manipulating the attackers into equipment connected.

What does AI bug mean

Google has given a clear idea of ​​what kind of weaknesses the researchers want to see. Imagine an attacker to unlock a smart door or use a hidden command to the Google Home, who summarizes someone’s email to Gmail and sends them to a third party. These are high-risk exploits that Google want to expose them before being used in the real world.

However, the company has made one thing clear, simply Mithun or any other AI model is not counted as a bug. What kind of material AI produces problems related to, such as abusive language or copyright content, should be directly reported within the product using its feedback tool. Google says it helps its AI security teams to retreat and improve the model in a more targeted manner.

How much can YOU earn

The biggest award – up to $ 20,000 – is reserved for vulnerable Google’s major products such as search, Gemini apps, Gmail and Drives. If a report stands out for its quality or originality, the bonus payment can push up to $ 30,000. The lower but still meaningful prizes will be given for defects discovered in other devices such as notebooks or experimental AI auxiliary jules.

According to Google, this step is part of a comprehensive effort to strengthen AI security as these systems are integrated deeply into its products. In fact, the company says that the researchers have already earned more than $ 430,000 by highlighting the AI-related risks in the last two years, before starting this official program.

Google launched Code Mender tool: AI Fixing AI

Along with the new reward program, Google also introduced a tool called Code Mender. It is an AI agent designed to automatically detect and fix security weaknesses in open-source software. So far, the codeomeder has helped patch more than 70 verified issues after being examined by human reviewers. The company says that such equipment shows how AI can also play a positive role in making technology more secure.