We’re entering a strange moment in digital life where seeing, hearing, or reading something is no longer enough to trust it. AI can now clone voices, mimic writing styles, generate fake videos, and imitate human behaviour with unsettling accuracy.
As these tools become more accessible, the real consumer challenge is no longer just spotting obvious scams, but figuring out how we determine who is actually real online.
The rise of AI-driven cyber threats hasn’t rendered passwords obsolete, but it has highlighted the critical need for multi-layered defence-in-depth to protect user accounts. We need to talk less about static credentials and focus on the deeper concept of identity integrity. Basically, the digital equivalent of the age-old principle of trusting what someone does rather than just what they say, notes Anna Collard, SVP, content strategy & CISO advisor for KnowBe4 Africa.
“The password isn’t dead, but the industry has moved toward a ‘Zero Trust’ model where a single set of credentials isn’t sufficient to grant access anymore,” says Collard. She points out that most modern breaches don’t involve sophisticated technical hacking, but rather the exploitation of exposed credentials. “Criminals obtain passwords through phishing or data leaks and simply log in as the user.
Grandmother’s logic in a digital world
To navigate the new era of deepfakes and AI clones, Collard suggests returning to a surprisingly low-tech source of wisdom. “Grandma would always tell you to trust actions over words,” she says. “That advice is surprisingly apt in modern cybersecurity. It is the shift from asking ‘What do you know?’ to ‘Who are you being?’
“Consider how you recognise a loved one on the phone,” Collard highlights. “You don’t just rely on the sound of their voice – which AI can now clone with terrifying accuracy – but on their cadence, their specific vocabulary, and your shared history.” If the person on the other end sounds right but speaks with an unfamiliar rhythm or lacks context, your behavioural alarm bells should ring.
The reCaptcha evolution
This concept of monitoring behaviour to verify identity is one many people already interact with, albeit in quite a basic form. “Most users are familiar with reCaptcha prompts – those website tasks asking you to identify things like traffic lights or buses,” says Collard. “While we think we’re proving our humanity by identifying photos, the system is actually watching the micro-movements of the mouse just as closely. The advent of reCaptcha pioneered using behavioural aspects to distinguish humans from bots, and modern behavioural telemetry is the sophisticated evolution of that principle.”
By moving from a single point of entry to a continuous stream of verification, organisations can ensure that trust is earned through every interaction, not just at the initial login.
Behavioural telemetry: The new signal of trust
By using technical signals – such as how a person types (keystroke dynamics), their typical work hours, and unique navigation patterns – organisations can verify identity in real-time. This behavioural telemetry creates a digital fingerprint derived from a user’s unique cognitive-motor pathways, making it significantly harder for automated bots or human adversaries to replicate.
“Even if a criminal possesses legitimate credentials and bypasses MFA, they (or their AI tool) won’t navigate your internal systems the way you do,” Collard points out. “An AI-driven impersonation might get the ‘what’ right, but often struggles with the ‘how’ and ‘when’. These subtle mismatches are where detection becomes possible. It’s less about catching a single anomaly and more about spotting patterns that don’t quite fit the human behind the purported identity.”
For instance, if an employee usually logs in at 9am from the office, it becomes easier for security systems to detect when the user’s location is unfamiliar or they’re logging in at 11pm at night. “Behavioural biometrics will determine what the employee’s familiar location is, the times they usually work and the way they access systems,” she says. “If any action doesn’t align with their role and past behaviour, it’s immediately flagged.”
This is where Zero Trust becomes real. “Trust needs to be continuously earned through behaviour,” Collard emphasises. “Even if the user appears to be logged in correctly, by analysing the interaction between their hardware and the application, these systems can identify when a device is being controlled remotely via a hijacked session.”
A holistic approach to Workforce Management
This shift is a core component of the Workforce Management framework KnowBe4 advocates for. In an environment where employees work alongside AI agents, protecting the identity of the employee becomes a core asset.
“Identity is the connective tissue between human behaviour and security risk,” states Collard.
She believes identity integrity is the lens through which organisations can assess susceptibility – who is being targeted – and real-time exposure. “It shifts security from simple awareness to adaptive risk management at the human layer,” she comments.
“Security is shifting towards continuous, context-aware verification, where behaviour is a signal of trust,” Collard concludes. “In this environment, detecting subtle deviations in user activity may prove far more valuable than the strength of any single credential. We are moving toward continuous authentication and digital familiarity, instead of checking your ID once at the door (the login), the system ‘watches’ you while you’re in the building to make sure you’re still acting like you.”
