Your AI Agent Went Rogue. Now What?
Most incident response plans assume a person is behind the mistake. Here's how to build one that holds up when an AI agent is the one acting.
1 min read
Heroic Technologies : Published Updated
Whatsapp has had a tough year from a security standpoint and has suffered losses in the size of its user base as a result.
To their credit, the company has been working hard to make their product more secure.
Recently they took another powerful step in that direction with the introduction of disappearing messages. These allow you to set a time delay that will cause your conversations with other users to have a pre-defined shelf life of either 24 hours or 90 days at your option.
This builds on the disappearing messages feature the company introduced last year which allowed users or admins to set messages to disappear after seven days.
“When enabled, all new one-on-one chats you or another person start will be set to disappear at your chosen duration, and we’ve added a new option when creating a group chat that lets you turn it on for groups you create. This new feature is optional and does not change or delete any of your existing chats.
For people who choose to switch on default disappearing messages, we will display a message in your chats that tells people this is the default you’ve chosen.
This makes clear it’s nothing personal – it’s a choice you’ve made about how you want to communicate with everyone on WhatsApp moving forward.
We believe disappearing messages along with end-to-end encryption are two crucial features that define what it means to be a private messaging service today, and bring us one step closer to the feeling of an in-personal conversation.”
These are good changes and they do indeed serve to mitigate the risks of Whatsapp users. Kudos to the company for continuing to improve their product and make it more secure.
Most incident response plans assume a person is behind the mistake. Here's how to build one that holds up when an AI agent is the one acting.
A due diligence checklist for any AI vendor: data handling, training-use terms, subprocessor disclosure, breach notification, and liability language.
Learn how to scope least-privilege access for AI agents: dedicated accounts, the identity, authority, autonomy, consequence framework, and review...
While “vishing” is by no means a new threat, it’s not something that has ever happened with sufficient frequency to get most people’s attention.
The UK’s National Cyber Security Centre (NCSC) has found over 4,000 retailers compromised by hackers exploiting known vulnerabilities on checkout...
Diana Lopera is a researcher for Trustwave Cybersecurity and has stumbled across something that’s one part interesting and one part disturbing.