Decision Drift: Why Good Judgement Isn't Enough Anymore
Published: July 2026
When reasonable decisions meet increasingly difficult environments
We trust familiar places. We follow established processes. We assume the next step is the right one. In AI-enabled, digitally mediated environments, "familiar" is getting harder to read.
Today, decisions move through platforms, systems, and recommendations whose legitimacy is difficult to judge from the inside. This article explains Decision Drift: why sound judgement can still lead people wrong, and offers a practical way to build in resistance to it.
Contents
What Decision Drift means
Decision Drift happens when the environment changes faster than the assumptions we bring to it.
Picture someone applying for a role on a reputable platform. A professional message arrives, followed by an interview invitation and a scheduling link. Nothing looks unusual, because nothing is unusual, at least not in the way they were trained to spot.
They didn't click a suspicious link. They followed a normal sequence of events on a platform they had every reason to trust.
That's the drift: the danger sits in the assumption that a trustworthy setting guarantees a trustworthy actor, not in the action itself.
Everything here looks legitimate. That is the point.
The myth of the suspicious link
Digital safety advice hasn't changed much in a decade. Check for poor grammar. Watch for unrealistic salaries. Be wary of unsolicited messages and urgency. Trust your instincts.
That advice assumes the con looks cheap. It rarely does anymore.
Grammar is polished, often AI-written on both sides of the interaction. Salaries are plausible. The message arrived through a platform you applied to directly. The link is framed as the next expected step, sent by someone posing as your recruiter. Urgency can be genuine. And your instincts, tuned to spot the old tells, read all of this as completely normal.
The old lesson was: don't trust strangers. The updated lesson is closer to this: don't let the credibility of the environment stand in for the credibility of the person acting inside it. The signal lives less in the message than in the route that got you there and the pace at which you're expected to act.
Trust as infrastructure
Modern life runs on platforms. Increasingly, we don't evaluate the person first: we evaluate the platform they showed up on. People rarely think "I trust this specific recruiter." They think "I found this through LinkedIn, Indeed, or a company career page," and let the platform's credibility stand in for the actor's.
For most people, that distinction (trusting the platform versus trusting the person) has become almost invisible. Trust has migrated into the infrastructure itself, which makes accountability harder to locate when something goes wrong: the platform carried the interaction, but nobody at the platform necessarily vouched for the actor.
That matters because when the environment supplies the trust, people stop pausing for independent confirmation. The setting becomes part of the evidence, which works fine when the setting is reliable, and fails badly when it lends credibility to the wrong actor.
Operationalising accountability
Before acting on a request that arrived through a platform, three questions are worth asking:
- Who owns this? Not the platform in the abstract: the specific person or team responsible for this interaction.
- Who owns the risk? If this goes wrong, whose problem is it: yours, the platform's, or nobody's?
- Who would help if it went wrong? Is there an actual escalation path, or just a support form?
If those questions can't be answered quickly, the uncertainty itself is the signal to slow down.
This isn't just an individual discipline. Organisations with strong policies and strong performance metrics can still miss what people on the ground are actually experiencing.
Frameworks like the NIST AI RMF and ISO/IEC 23894 give organisations solid scaffolding for AI risk. But both govern systems, not the moment a person quietly absorbs risk because the environment made it feel safe to. That gap is where Decision Drift lives, and it's why accountability has to work at the human level, not just the policy level. For a practical crosswalk between the two, see our Prime 7 Gap Analysis: Human Risk in AI Governance.
The pilot's checklist
Pilots use checklists because complex environments punish improvisation. The same logic applies here. Before proceeding on a decision that feels routine but originated through a platform or automated recommendation, ask:
- Can I independently verify the organisation, separate from the platform that introduced it?
- Am I being moved away from a trusted environment faster than the situation actually requires?
- What evidence do I have that this specific destination (not the platform, the destination) is genuine?
- If this goes wrong, do I already know who's accountable?
The checklist doesn't need to be exhaustive. It needs to create a pause at the moment the context shifts.
Deliberate interruption points
Digital resilience works through planned pauses, not blanket suspicion, a deliberate moment built into the process right when a person would otherwise ask: "What am I assuming about this environment that I haven't actually verified?"
Complex environments don't need better people. They need better pauses, placed where trust is assumed rather than earned.
Final thoughts
We spent years teaching people to spot the suspicious link. What caught up with us were links that looked exactly like the next ordinary step in an ordinary process.
People haven't gotten worse at judgement. The environments they're judging have gotten harder to read, and scams no longer need to build their own credibility: they can borrow it from the platform underneath them.
The question worth sitting with: if modern platforms function as custodians of trust, what responsibility comes with that role, and who's actually holding it?
Q&A
How is Decision Drift different from bad judgement?
It isn't a failure of judgement at all. It's what happens when reasonable people make reasonable decisions inside environments that have quietly gotten harder to verify.
What's the first move when a decision feels normal but something's still off?
Pause. Separate the platform's credibility from the actor's, and get independent verification before taking the next step.
Who's most at risk?
Anyone relying on platforms for work, opportunity, or communication: in practice, that means almost everyone, but especially job seekers, leaders approving vendor or partner requests, and teams under time pressure.
What can organisations actually do about it?
Make accountability visible and specific, build deliberate pauses into fast-moving workflows, and give people a real, fast way to verify who owns a process before they act on it.
This article draws on research into trust and trustworthiness in AI governance, and on emerging work examining how organisations experience decision drift and decision leakage in practice.
About CKC Cares
CKC Cares helps leaders build practical, human-centred capability for digital resilience and responsible technology, translating complex safety thinking into habits usable on ordinary workdays.
Disclaimer: This article offers general guidance on leadership practice and digital resilience. It is not legal advice, medical advice, or a substitute for a specialist assessment of your specific systems, policies, and organisational context. For guidance tailored to your situation, book a training, workshop, coaching, or advisory session at theclarityline.co.uk. CKC Cares does not assume responsibility for decisions made based on this information.
