A couple living in a California town near San Francisco told local media they experienced "sheer terror" over the weekend when a Nest security camera atop their family's television issued a realistic-sounding warning of missiles heading to the United States from North Korea.

The couple and their frightened child eventually figured out they had been targeted by a hacker who got control of speakers built into the camera, which is equipped for two-way conversations.

Nest, which is owned by Google-parent Alphabet, told AFP that incidents of commandeered camera control in recent months were the result of hackers using passwords stolen from other online venues.

"Nest was not breached," Google said in a statement.

"These recent reports are based on customers using compromised passwords - exposed through breaches on other websites."

Reported incidents involving Nest cameras in the US in the past three months include a hacker threatening to snatch a baby and a seemingly well-intended hacker telling someone that his data was exposed.

Nest camera users were urged to prevent such invasions by implementing two-factor verification, meaning that a second step such as entering a code sent via text message is needed along with a password to get into an account.

Nest sifts through stolen data dumped online by hackers to check whether email addresses and passwords match those used for accounts at the smart home device company.

Account holders are prompted to change passwords when matches are found, but the massive amount of stolen data posted online by hackers can make the process slow.

People can check online at sites such as www.haveibeenpwned.com to see whether their email addresses have been found in troves of stolen data.