It said it had fixed the flaw and informed the relevant law enforcement officials about it, according to a statement from Guy Rosen, vice-president of Product Management.
Facebook got rid of their CISO position months back and folded their security function, opting to put security staff into each area of business. A bold move.— Kevin Beaumont (@GossiTheDog) September 28, 2018
The news came the same day that a white hat from Taiwan said he would delete Zuckerberg's account on Sunday. According to the Daily Mirror, Chang Chi-yuan said he would delete the account while others watched.
Rosen said the flaw had been found on 25 September. "[It] allowed them [the attackers] to steal Facebook access tokens which they could then use to take over people’s accounts. Access tokens are the equivalent of digital keys that keep people logged in to Facebook so they don’t need to re-enter their password every time they use the app," he said.
"As a result, around 90 million people will now have to log back in to Facebook, or any of their apps that use Facebook Login. After they have logged back in, people will get a notification at the top of their News Feed explaining what happened."
Rosen said the flaw had been caused by a change made to the video uploading feature in July 2017.
"The attackers not only needed to find this vulnerability and use it to get an access token, they then had to pivot from that account to others to steal more tokens," he said.
Commenting on the incident, Matt Chiodi, vice-president of Cloud Security at RedLock, said the vulnerabilities underscored the level of persistence on the part of attackers.
"If there’s a high enough value target, they will get in sooner or later," he said. "Remember that Facebook today employs over 10,000 cyber security professionals.
"No system or application is 100% secure. What's most intriguing is that despite a formal bug bounty program, the vulnerability has been present in Facebook code since July of 2017. It’s hard to believe that a vulnerability of this size would persist this long undetected.”
Rahul Kashyap, chief executive of security outfit Awake Security, said the immediate challenge was for Facebook to identify the accounts that had been touched and those which had been compromised.
"The 50 million number could change as we often have seen with past breaches. But it is quite likely a subset of those were specifically taken over," he said.
In the wake of the #Facebook news, here's a timeline (dating all the way back to 2003) of the other times #Facebook did bad things. #sorrynotsorry https://t.co/3gg6JBVuty— Privacy International (@privacyint) September 28, 2018
We're sure there are ones we're missing so please share links and we'll update the timeline! pic.twitter.com/0W01Zc2QlQ
“What will be revealing is whether there is a pattern to whose accounts were being targeted, and whether that pattern will help reveal the identity of the attackers.
"Facebook knows what it knows now, but there’s always the possibility that attackers were able to get to more information. The large numbers in this breach could just be a decoy if threat actors were targeting specific individuals.
"This part of Facebook’s statement says it all: 'Since we’ve only just started our investigation, we have yet to determine whether these accounts were misused or any information accessed. We also don’t know who’s behind these attacks or where they’re based. We’re working hard to better understand these details – and we will update this post when we have more information, or if the facts change'."