It was only a matter of time before hackers started using artificial intelligence to attack artificial intelligence—and now that time has arrived. A new research breakthrough has made AI prompt injection attacks faster, easier, and scarily effective, even against supposedly secure systems like Google’s Gemini.

Prompt injection attacks have been one of the most reliable ways to manipulate large language models (LLMs). By sneaking malicious instructions into the text AI reads—like a comment in a block of code or hidden text on a webpage—attackers can get the model to ignore its original rules.

That could mean leaking private data, delivering wrong answers, or carrying out other unintended behaviors. The catch, though, is that prompt injection attacks typically require a lot of manual trial and error to get right, especially for closed-weight models like GPT-4 or Gemini, where developers can’t see the underlying code or training data.

But a new technique called Fun-Tuning changes that. Developed by a team of university researchers, this method uses Google’s own fine-tuning API for Gemini to craft high-success-rate prompt injections—automatically. The researcher’s findings are currently available in a preprint report.

By abusing Gemini’s training interface, Fun-Tuning figures out the best “prefixes” and “suffixes” to wrap around an attacker’s malicious prompt, dramatically increasing the chances that it’ll be followed. And the results speak for themselves.

In testing, Fun-Tuning achieved up to 82 percent success rates on some Gemini models, compared to under 30 percent with traditional attacks. It works by exploiting subtle clues in the fine-tuning process—like how the model reacts to training errors—and turning them into feedback that sharpens the attack. Think of it as an AI-guided missile system for prompt injection.

Even more troubling, attacks developed for one version of Gemini transferred easily to others. This means a single attacker could potentially develop one successful prompt and deploy it across multiple platforms. And since Google offers this fine-tuning API for free, the cost of mounting such an attack is as low as $10 in compute time.

Google has acknowledged the threat but hasn’t commented on whether it plans to change its fine-tuning features. The researchers behind Fun-Tuning warn that defending against this kind of attack isn’t simple—removing key data from the training process would make the tool less useful for developers. But leaving it in makes it easier for attackers to exploit.

One thing is certain, though. AI prompt injection attacks like this are a sign that the game has entered a new phase—where AI isn’t just the target, but also the weapon.

The post Hackers are now using AI to break AI – and it’s working appeared first on BGR.

Today's Top Deals

1. Today’s deals: $169 AirPods Pro 2, $100 off Breville air fryers, $40 Echo Dot, $250 Sony XM5 headphones, more
2. Here are our favorite Magenta Status perks you can only get from T-Mobile
3. Mother’s Day gift ideas 2024: Thoughtful gifts mom will never forget
4. Amazon Big Spring Sale: Apple deals, gaming laptops, Crest 3D Whitestrips, Samsung deals, more

Hackers are now using AI to break AI – and it’s working originally appeared on BGR.com on Sat, 29 Mar 2025 at 09:01:00 EDT. Please see our terms for use of feeds.

Tesla CEO Elon Musk's losing streak continued on Wednesday.

Hours after Republicans in Wisconsin lost a state Supreme Court race in Wisconsin where Musk spent tens of millions of dollars of his own money, shares of his flagship electric car company fell upon news that it had badly missed Wall Street estimates for the number of vehicles it had delivered in the first quarter of 2025.

As reported by Market Watch, Tesla revealed that "it delivered more than 336,000 vehicles during the first quarter, down from 387,000 deliveries a year ago."

To put this number into context, Market Watch notes that it "was well below the average analyst estimate compiled by FactSet of 404,000 vehicles" and it "was also lower than the 'whisper' number, or what Wall Street was quietly expecting, with even the most bullish analyst projecting a number in the 360,000 range."

ALSO READ: 'Mad king': Trump baffles observers by suggesting he'll tariff illegal drugs

Although Tesla blamed the disappointing deliveries on what Market Watch describes as "the changeover to refreshed Model Y production lines" that "led to several weeks of lost production," it's also a reality that the company has been the target of a concerted boycott campaign led by progressive groups to protest Musk's role as the head of President Donald Trump's Department of Government Efficiency.

Sales of Tesla in Europe have fallen off a cliff in recent months, as car buyers on that continent have revolted against Musk's support for far-right political parties such as Germany's Alternative für Deutschland.