Canines are known for their incredibly accurate sense of smell but it turns out that drug detection dogs may be all bark and no bite.

An Australian survey of law enforcement K-9 units revealed that drug sniffing dogs Down Under signaled incorrectly 75 percent of the time. The findings have led to calls from Australian lawmakers to cease the deployment of drug sniffing dogs outside music festivals, where many of the searches took place. 

The analysis, reported by SBS News, examined more than 94,000 searches initiated by a drug dog alert over a ten-year period. Of those cases, roughly 71,000 ended with no drugs found.

Greens MP Cate Faehrmann argued that using dogs to initiate searches is unnecessary.

“Unequivocally, drug dogs don’t stop people taking drugs, they just lead to riskier behavior and sometimes that riskier behavior can have fatal outcomes,” she said outside of a recent Sydney concert, according to SBS News.

The Australian study is not the first time the accuracy of drug dogs have been called into question. A survey conducted by the Chicago Tribune found that K-9s in suburban Chicago indicated correctly 44 percent of the time.

While it may be easy to blame the dogs for their confusion, science suggests it has more to do with their human companions. A 2011 study found that the beliefs of dog handlers influenced the actions of the scent detection canines; i.e., if the handling officer believes a suspect has drugs, their K-9 officer likely will too.

Despite the track record, drug detection dogs are used widely by law enforcement, airport security, and even cruise lines hoping to prevent drugs from being brought aboard. The Supreme Court has said that an alert from a dog does not constitute a search, but does offer probable cause for a search to be executed. 

The Australian findings may lead to renewed interest in the place of drug sniffing dogs in law enforcement. However, unless a landmark case shifts the opinion of the Supreme Court, the status quo will likely remain—at least in the U.S.