Latest / Columbus' Morning News / What is Google's "Zero Day" discovery?
Transcript
- 0:00Speaker 1: All right, we'll switch over to the Legacy Retirement Group
- 0:02Speaker 1: dot Com phone line. Say good morning to Mike Debuski,
- 0:05Speaker 1: ABC News Technology reporter. It is Tech Tuesday. In what
- 0:09Speaker 1: is Google's zero day discovery?
- 0:12Speaker 2: Yes, so this is a major cybersecurity story, guys. According
- 0:17Speaker 2: to Google on Monday, they said that they've identified a
- 0:20Speaker 2: group of hackers who were potentially trying to exploit a
- 0:23Speaker 2: flaw in two factor authentication. Now, right up front, I
- 0:27Speaker 2: want to say, we don't know a lot of stuff here.
- 0:30Speaker 2: Google did not identify this criminal hacking group. We don't
- 0:33Speaker 2: know what companies the hackers were targeting. We don't know
- 0:36Speaker 2: when this happened or how it was stopped. But Google
- 0:39Speaker 2: does say it's not currently a threat anymore. And we
- 0:43Speaker 2: also don't know what type of artificial intelligence was used
- 0:46Speaker 2: in this hacking effort. And that's an important piece here, guys.
- 0:50Speaker 2: If you think back a couple of weeks, Anthropic made
- 0:52Speaker 2: this big announcement about Mythos, which is apparently this AI
- 0:56Speaker 2: model that was just so powerful. Anthropic chose not to
- 0:59Speaker 2: release it. In large part, they decided to do that
- 1:01Speaker 2: because they were concerned that hackers would be able to
- 1:04Speaker 2: use Mythos to comb through huge complex security systems around
- 1:09Speaker 2: the world, identifying vulnerabilities in major companies, government organizations, whatever
- 1:13Speaker 2: it happens to be. On Monday, Google Threat Intelligence Group
- 1:17Speaker 2: said that this was basically something that was happening, right,
- 1:20Speaker 2: that hackers were using some kind of AI to identify
- 1:25Speaker 2: flaws in a system administration tool related to two factor authentication.
- 1:30Speaker 2: And then you think about how many things are protected
- 1:33Speaker 2: by two factor authentication. That is a real concern, right,
- 1:37Speaker 2: It's kind of an annoying thing, but it's also thought
- 1:39Speaker 2: to be one of the better digital security measures that
- 1:42Speaker 2: are out there. This flaw is what's known as a
- 1:45Speaker 2: zero day vulnerability, which is a very scary terminology identifying
- 1:50Speaker 2: what cybersecurity experts really have nightmares about. Guys, there are
- 1:53Speaker 2: these vulnerabilities in these very complex sort of scores of
- 1:57Speaker 2: code that developers are not aware of. Every piece of
- 2:00Speaker 2: code has, you know, some kind of flaw in it eventually,
- 2:04Speaker 2: and these are the flaws that are really dangerous because
- 2:07Speaker 2: the developers are not aware of them. Right, they can't
- 2:10Speaker 2: patch them, they can't work on fixes. You don't know
- 2:13Speaker 2: what you don't know, and that's exactly why these things
- 2:16Speaker 2: are so dangerous and when you apply that to something
- 2:19Speaker 2: like two factor authentication, the risk here is pretty big.
- 2:21Speaker 1: Yeah, I thought two factor authentication was sort of the
- 2:23Speaker 1: gold standard. I mean, you go to a certain app
- 2:26Speaker 1: or a website and then you click on a hey
- 2:28Speaker 1: send me the code sends you the code to your phone,
- 2:31Speaker 1: you type the code back in. So how would they
- 2:33Speaker 1: get access? Would they get access from the program that
- 2:38Speaker 1: would send the code out to your additional mobile device?
- 2:42Speaker 2: So from what we understand, this was a hack that
- 2:45Speaker 2: basically allowed them to bypass two factor authentication to get
- 2:49Speaker 2: around that particular wall. However, they also needed like legitimate credential.
- 2:54Speaker 2: So you think, if you're plugging into I don't know
- 2:56Speaker 2: how tvns, you know, computer systems work or anything like that,
- 2:59Speaker 2: but an email address and a pass word and then
- 3:01Speaker 2: it sends you a two factor authentication. That's what they
- 3:04Speaker 2: do here at ABC, And like, okay, that kind of
- 3:07Speaker 2: makes some degree of sense. These hackers, if they were
- 3:10Speaker 2: to have gotten into any of these systems that they
- 3:12Speaker 2: were targeting, would still need like legitimate email addresses and
- 3:16Speaker 2: passwords and stuff like that. You know, they could conceivably
- 3:19Speaker 2: have gone and gotten that stuff, right, That stuff leaks
- 3:22Speaker 2: all the time. In data breaches and whatnot. And and
- 3:24Speaker 2: that's a big concern something that two factor authentication was
- 3:27Speaker 2: supposed to guard against.
- 3:29Speaker 1: If they have a way around that, well that's that's
- 3:31Speaker 1: really worrying for a lot of time. So where we
- 3:33Speaker 1: see three factor authentication. I don't have a good answer
- 3:37Speaker 1: to that. That's a good question. Where do you go
- 3:39Speaker 1: after this? I'm kidding, half kidding. I don't know, Like,
- 3:41Speaker 1: what do you like? But ours go to eleven? I
- 3:43Speaker 1: don't know how you go up from two factor speaking
- 3:46Speaker 1: with Michael the Busky, ABC News Technology reporter Tech Tuesday.
- 3:50Speaker 1: So the another lawsuit on chat GPT from the family
- 3:56Speaker 1: of the FSU shooting victim. The you're starting to see
- 3:59Speaker 1: these now, you know, chat GPT could have prevented some
- 4:03Speaker 1: of these shootings and they did not.
- 4:05Speaker 2: That's what this lawsuit is alleging. Yeah, so Florida State
- 4:08Speaker 2: University was the scene of a deadly mass shooting last year.
- 4:13Speaker 2: Tiru Chaba was one of the people killed. It's his
- 4:16Speaker 2: widow now lodging this lawsuit against open Ai, Chat GPT
- 4:20Speaker 2: and Phoenix Ickner, who is the accused gunman in this case.
- 4:24Speaker 2: According to this lawsuit, chat GPT provided Ickner with input
- 4:27Speaker 2: and information when planning this shooting. The lawsuit sites thousands
- 4:31Speaker 2: of chats that Ackner allegedly had with this AI chat
- 4:33Speaker 2: bought about everything from how certain guns work to what
- 4:36Speaker 2: types of shootings get the most media attention. He was
- 4:39Speaker 2: researching previous mass shootings apparently as well as Hitler fascism,
- 4:43Speaker 2: Christian nationalism. This lawsuit also alleges that OpenAI failed to
- 4:47Speaker 2: create a product that would refrain from participating in these
- 4:50Speaker 2: discussions and that, according to this lawsuit, amounted to it
- 4:54Speaker 2: co conspiring with Ickner to commit these crimes. They are
- 4:58Speaker 2: looking for an unspecified amount of damages here. It is
- 5:02Speaker 2: an eight count complaint that accuses these defendants Chat, GPT
- 5:06Speaker 2: and open Ai of wrongful death, negligence, gross negligence, and
- 5:11Speaker 2: it accuses the AI company specifically here. Ickner's also mentioned
- 5:14Speaker 2: in this lawsuit of strict product liability, so things like
- 5:19Speaker 2: defective design, negligent design. This is interesting, guys, because it
- 5:24Speaker 2: is kind of reminiscent of the lawsuits against Meta and
- 5:27Speaker 2: the social media companies, not necessarily targeting the content that
- 5:30Speaker 2: is on these platforms, but rather the decision making in
- 5:33Speaker 2: Silicon Valley that led to certain features that are harmful.
- 5:37Speaker 1: So you would think, Mike, that there would be keywords
- 5:39Speaker 1: or key phrases that would trigger chat GPT or are
- 5:42Speaker 1: these AI chatbots to actually flag it and send it
- 5:46Speaker 1: to an actual human that could look at it and
- 5:48Speaker 1: determine if it was a thread.
- 5:49Speaker 2: Or not two a degree. Guys, there already are these flags,
- 5:52Speaker 2: and it varies based on what chatbot you're using, or
- 5:56Speaker 2: what company it belongs to, or in those types of things.
- 5:58Speaker 2: But obviously, you know, you know, these these systems are
- 6:03Speaker 2: coming under a lot of scrutiny because these you know,
- 6:07Speaker 2: kind of mass shooting incidents, you know, incidents of violence
- 6:11Speaker 2: and other things are you know, found to have at
- 6:14Speaker 2: least you know, had their origins in some way on
- 6:17Speaker 2: these platforms. People having conversations with chatbots and whatever that
- 6:22Speaker 2: you know, either we're not flagged or were not considered
- 6:25Speaker 2: serious enough to be flagged, you know, something like that. Now,
- 6:30Speaker 2: chat GPT and open Ai, for their part, have responded
- 6:33Speaker 2: to all this. They call the shooting at f SU
- 6:35Speaker 2: a tragedy, but that chat GPT is not responsible for
- 6:37Speaker 2: the crime. It said they did not encourage or promote
- 6:40Speaker 2: illegal or harmful activity, and they made the argument that
- 6:43Speaker 2: these you know materials, the information that chat GPT was
- 6:47Speaker 2: able to provide to Ickner is freely available across the
- 6:51Speaker 2: Internet right It's found broadly across public sources, according to
- 6:54Speaker 2: their statements. So it's going to be interesting to see
- 6:56Speaker 2: how this progresses because there's a lot of thornyiness here