Hypothetically to guard against double-stranded RNAs trying to incorporate themselves into our DNA (most of humans’ normal RNAs are single-stranded), we’ve all got an enzyme that cuts these dsRNAs into pieces. It’s aptly named Dicer. Dicer chops up these pieces into small bits of double-stranded RNA, now called small interfering RNAs (siRNAs). The siRNAs then bind to mRNAs (messenger RNAs), generally of the same sequence as the dsRNA; this then activates *another* enzyme complex that detects double-stranded mRNA, and breaks it down.
Theoretical translation? If you’ve got a gene that’s malfunctioning, or a virus that’s multiplying, or a bacteria that’s making toxic proteins, you could (again in theory), develop a double-stranded RNA that matches with the malfunctioning or deleterious mRNA, and the body’s own defense proteins would chop it up, meaning the harmful proteins would never be created in the first place.
Theory doesn’t always work out in the slow and methodical world of bench research, but there are already RNAi databases, companies sprouting up that work solely on RNAi techniques and research, and RNAi researchers around the world.