theguardian.com via Reddit

PR Pros Expose Clients Pitching Old Automation as AI

generative ai enterprise ai ai-business marketing ai-ethics

Key insights

  • Roughly half of AI-related press releases are sent under client pressure, per named PR executives interviewed by The Guardian.
  • Rebranded tools include regression models, statistical analyses, and Excel automation, none of which meet standard AI definitions.
  • The investigation is among the first to source AI hype manufacturing at the PR layer, tracing demand to non-tech clients across industries.

Why this matters

The scale of documented AI washing directly degrades the signal quality of AI adoption metrics that investors, analysts, and founders rely on to benchmark real market penetration. For technical leaders auditing competitive landscapes, a material share of competitor AI announcements likely describe decade-old statistical tooling, meaning competitive intelligence built on press coverage is systematically inflated. Regulators at the FTC and EU AI Office building governance frameworks around self-reported AI usage will need to contend with systematic misclassification originating at the press-release layer, not just at the product layer.

Summary

PR professionals are on the record saying roughly half of the AI-related press releases they send go out under client pressure, pitching tools that aren't AI by any technical definition. The Guardian spoke with named communications executives who describe regression models, statistical analyses, and Excel macros being repackaged as AI products. One called the trend 'completely unhinged.' Clients include shoe brands and property companies, making this a cross-industry pattern rather than a tech-sector quirk. Essentially: non-tech brands are the primary demand source, with PR firms serving as reluctant distributors of inflated claims. - Roughly half of AI-labeled press releases are sent under client duress, per interviewees. - Rebranded tools include regression models, statistical analyses, and Excel-based automation, none qualifying as AI under any standard definition. - This is one of the first sourced investigations tracing AI hype to the PR supply chain rather than to tech companies themselves. AI hype isn't just a tech industry export; it's being manufactured by mainstream brands across every sector, afraid of being seen as left behind.

Potential risks and opportunities

Risks

  • Investors using AI adoption announcements as a productivity-gain proxy may be systematically overvaluing non-tech companies that have made no real AI investment, with correction risk rising as auditors and analysts tighten definitions.
  • The FTC or EU AI Office could pursue enforcement actions against brands that made materially false AI claims in earnings disclosures or regulated marketing materials, with PR firms potentially named as knowing distributors.
  • PR agencies that distributed fabricated AI claims face reputational and legal exposure if client-facing AI announcements are scrutinized in shareholder suits or regulatory proceedings within the next 12 to 18 months.

Opportunities

  • Third-party AI audit and transparency vendors (Credo AI, Holistic AI) have a clear opening to sell pre-publication claim verification services to brands seeking credibility before making public AI statements.
  • Investment analysts and research firms that build systematic AI-claim verification databases, cross-referencing press releases against technical disclosures, can monetize the growing gap between claimed and verified AI adoption.
  • Enterprise PR software platforms (Muck Rack, Cision) could differentiate on compliance tooling that flags AI-related claims against FTC guidance before release, turning regulatory risk into a product feature.

What we don't know yet

  • Whether any of the named PR firms have internal policies for verifying AI claims before distribution, or have declined client work on these grounds.
  • What share of AI-related references in non-tech company earnings calls describe similarly repackaged legacy automation, a dataset this investigation did not attempt to quantify.
  • Whether The Guardian will publish the underlying firm or client names, which would allow researchers to cross-reference against SEC filings or FTC disclosure records.