San Francisco prosecutors will use artificial intelligence technology to try to reduce racial bias on whether to charge suspects with a crime.
Prosecutors have teamed up with Stanford Computational Policy Lab, who have developed a programme which takes out the suspect's name, race, hair and eye colour from police reports.
The world-leading technology would remove any possible information that could sway prosecutors before they receive the report, leaving them with the facts of the case.
By doing this, district attorney George Gascon said he hopes to take "race out of the equation" when it comes to criminal disputes.
In a recent interview with US media, Mr Gascon said: "If we can take racial bias out of our system or reduce it significantly, we can be a much better nation.
“It’s not only improving our system but creating a model that can be used elsewhere.”
There are plans for the technology to be used from July 1, but Mr Gascon said he hopes other forces across the US would be able to use the technology soon if it proves successful.
Researchers at Stanford Computational Policy Lab who have been working on the new technology. They said that over the last few decades, "intuitive judgments are often inferior to statistically informed assessments".
The researchers admitted there is "no single metric for determining whether and algorithm is fair", but "there are still many jurisdictions that use algorithmic tools to predict arrests for any offense, including petty street crimes, potentially introducing racial bias into their assessments".
"Alongside more comprehensive reforms, algorithmic risk assessments have the potential to make the criminal justice system more equitable.
"But they must be developed, evaluated, and implemented with care that reflects the seriousness of the decisions they help guide."
A report commissioned by the San Francisco distinct attorney found "substantial racial ethnic disparities in criminal justice outcomes."
While African-Americans accounted for just six per cent of the county's population, the same demographic represented 41 per cent of arrests between 2008 and 2014.