I was surprised this week with the news that several Twitter celebrity accounts were compromised and used in a common Cryptocurrency scam. Speculation on how this happened, given related postings of Twitter’s internal user control panel, would indicate the cause could have been someone hacking into a Twitter employee’s corporate account to gain access to this tool.
The scam follows what has become a standard going back to even before the Internet. It asks you to send money to get money. Defending against this scam should be taught in grade school, given how common it is because similar scams have used snail mail and hard-line phone transport methods.
But one of the applications for AI is identity assurance. AI Identity assurance was demonstrated by Cylance, a division of Blackberry, some time ago. The AI learns how you type, where you type from, and what you type about, and if it sees something that doesn’t match your profile, it locks the user out and forces a reauthentication. Had Twitter applied this technology, they wouldn’t have been embarrassed, but let’s go broader with this because I think a Cylance-like AI could fix much of the problem with social media.
Now the reason Blackberry’s Cylance division was front of mind is that they are rumored to be in the process of being acquired. I think it is this Cylance capability that could be the reason, because it could result in a far safer future Social Media service. Still, it could also far better protect executives from mistakes.
How Applied AI Could Fix Social Media
If you think about social media broadly, there are a ton of problems staring with hostile third parties creating fake accounts or taking over real ones and ranging to people posting things that cost them their jobs and reputations.
I subscribe to an AI tool from Experion, for instance, that will alert me if I’ve posted something that could damage my reputation. It is helpful as a learning tool. But given I get the alert well after I’ve posted, should I do something foolish, chances are it will just rub salt into the wound because, by the time I get the alert, everyone else would have already seen my foolishness and the damage would have been done.
But what if the tool was embedded in either the service or generic Social Media tool like Hootsuite? Or, more broadly, in an email tool like Outlook, which is growing to become a social media tool? If you could lock your social media accounts to the tool and prevent a third party, including someone working inside of a social media company, from gaining access to your account, couldn’t you at least protect yourself?
Let’s talk about the benefits of such a tool.
Benefits Of Social Media Personal AI
Now the most significant benefit would be scanning a post before you make it for inappropriate language or objectionable material. You could still post, but at least you’d know of the risk you were taking. For instance, a few years back, an executive traveling on vacation to Africa made an inappropriate joke on social media and later found out that she’d been fired because of it. Now, if she’d received an alert before posting that said, “you may want to rethink this, there is an 80% chance that this will upset your friends and employer, and a better than 30% chance that it will cost you your job.” She wouldn’t have posted and wouldn’t have been fired.
Now concerning the current breach that hit folks like Elon Musk, Bill Gated, and President Obama. A timely alert to any of them that would have resulted in a call to Jack Dorsey would have been more timely in his response. Besides, properly applied, the technology would have immediately seen that the Tweet was coming from a location that wasn’t near where the celebrities were and contained content that was unusual for them freezing the accounts and deleting the Tweet almost immediately.
Coming up to the November elections, there are a lot of concerns a hostile State will now seize some of these accounts to change an election outcome. A false Tweet from the President or challenger about dropping from the race on election day, or admitting wrongdoing, or encouraging their followers to vote for the other guy could have a material impact on the race and likely sink the related Social network.
In short, the very survival of these social networks could depend on deploying a tool like this. But deployed as part of an email or social media client could at least protect your folks from most of these problems.
Wrapping Up: The Critical Need For AI In Social Media
When the news broke about the breach, I’m sure a lot of Corporate Communications folks suddenly realized they had a new critical exposure. What if a hostile entity took over an executive’s account and posted something racially insensitive, announced a fake financial event (we are merging with a competitor – going out of business, etc.), advocated for a competing offering, or endorsed a questionable political candidate. That’s on top of usual concern of their just not thinking before posting.
This Twitter breach is a warning of what is to come. It is a showcase of why we need an AI-focused tightly on reputation and identity protection as soon as possible – as I expect threats will arise shortly.