AI slop is eating the world. Trust is the first casualty | TechRadar

If you're a small B2B SaaS founder using AI to draft your outbound emails, your investor updates, or your website copy, the ground is shifting under you — and it isn't about getting caught. It's about being believed.

A new piece from Actian CTO Emma McGrattan lays out why: as AI-generated content floods every channel, the audiences on the other end have started reflexively discounting anything that smells assembled rather than thought through. That reflex doesn't stay confined to LinkedIn posts and pitch decks — it spreads to every claim a company makes about itself, including the security and compliance claims on your website.

What actually happened

McGrattan's argument is that AI slop isn't just an annoyance, it's a slow-motion trust collapse. The mechanism is simple: putting your name on something used to be a signal that a person had actually thought about it. When AI generates the thinking too, that signal goes hollow — but audiences keep expecting it, and they notice when it's missing.

"It is a denial of service attack. Done at scale, they do not just waste time. They degrade the channel for everyone trying to use it honestly."
— TechRadar

She points to real, measurable fallout: response rates to AI-drafted cold outreach have collapsed, and investors reviewing pitch decks are getting faster at spotting when a strategic narrative was assembled by a prompt rather than discovered by actual work. Even the internet's collective knowledge base is shrinking under the same pressure — new questions posted on Stack Overflow are down almost 80% year over year, because fewer people bother writing out a real question when AI can paper over the gap instead.

Why this matters to a small B2B SaaS

Picture a 15-person vertical SaaS company mid-way through a $40,000 ARR enterprise deal. Their website's security page reads the way most do: "we take your data seriously," "enterprise-grade encryption," a vague nod to compliance. A year ago, a buyer's security reviewer might have skimmed past that boilerplate. Today, that same reviewer has spent months getting burned by AI-assembled decks and templated vendor questionnaires, and their default posture toward unverified claims has shifted from curiosity to suspicion. Generic assurance language now reads as a red flag instead of reassurance — and the deal stalls not because of an actual vulnerability, but because nothing on the page proves the claim is real.

That's the same "borrowed credibility" problem McGrattan describes, just relocated to the security page instead of the pitch deck. A claim with no evidence behind it costs you trust in 2026 in a way it didn't three years ago.

What to actually do about it

• Replace vague security assurances ("bank-level encryption," "we take security seriously") with specific, checkable facts: what's encrypted, what auth is enforced, what's publicly exposed.
• Get an outside, evidence-based read on what your company actually looks like from the outside — not what your marketing copy claims — before a buyer's security team does it for you.
• Put your real posture in writing somewhere a prospect can find it before they ask, so the first data point they get isn't a canned line that reads like everyone else's.
• Have whoever answers security questionnaires speak from actual knowledge of your systems, not a templated response bank — the same "did a person actually think about this" test buyers are now applying everywhere else.
• Revisit this every few months. Posture drifts as you ship; a claim that was true at signup can quietly stop being true a year later.

This is exactly the gap OrangeStealth's External Security Posture Assessment is built to close — a passive, external-only look at what a prospect or investor would actually find if they checked, so your security claims are backed by evidence instead of the same boilerplate everyone else is now discounting.