‘AI washing’: Why companies are racing to dress up automation as artificial intelligence

The trend: AI on every press release

Across the UK, PR teams are slipping the term artificial intelligence into announcements that, at their core, describe fairly ordinary rule‑based scripts or simple data pipelines. The pattern has been called “AI washing” – a marketing sleight of hand that promises futuristic tech while delivering incremental efficiency gains.

The phenomenon shows up in quarterly reports, hiring ads, and even in the branding of legacy software products. A recent survey of 150 mid‑size firms found that 62 % had added “AI‑enabled” to at least one product description in the past six months, even though the underlying code had not changed.

Evidence on the ground

1. Press releases that stretch definitions – Companies such as a logistics provider that advertises an “AI‑driven routing engine” when the system merely runs a deterministic Dijkstra algorithm. The language is identical to that used by startups whose models actually learn from data.

2. Recruitment signals – Job boards are flooded with titles like “AI Engineer” or “Machine‑Learning Analyst” for roles that involve writing SQL queries or configuring off‑the‑shelf RPA bots. Candidates report that the interview focus is on data‑pipeline hygiene rather than model development.

3. Investor communications – Venture capital decks now feature AI buzzwords on slides that still discuss “process automation” and “cost reduction”. Analysts note a spike in valuation multiples for firms that can claim any AI component, regardless of depth.

4. PR firm testimonies – Several UK‑based communications agencies have confirmed that clients explicitly request “AI language” to differentiate themselves in crowded markets. The request often comes from senior leadership rather than technical staff.

Why the rush?

• Market pressure – AI has become a shorthand for innovation. When competitors tout “AI‑powered insights”, a firm that does not appear on the same radar risks being labeled as stagnant.
• Talent attraction – The term draws engineers who want to work on cutting‑edge problems, even if the day‑to‑day work is more about integration than research.
• Funding incentives – Investors, still enamored with the hype cycle, are more willing to allocate capital to companies that can point to an AI roadmap, however vague.

Counter‑perspectives: Is the hype harmful?

Credibility erosion

Critics argue that AI washing dilutes the term’s meaning, making it harder for genuine breakthroughs to stand out. When every vendor claims AI, buyers become skeptical, leading to longer sales cycles and a “trust deficit”.

Regulatory risk

The UK’s upcoming AI governance framework may require firms to substantiate claims about automated decision‑making. Overstating capabilities could trigger compliance reviews or even legal challenges if consumers are misled.

Opportunity cost

Resources spent on rebranding could be redirected toward actual model development. Small teams that focus on building modest but real machine‑learning pipelines often achieve measurable ROI, such as a 12 % reduction in churn for a subscription service that used a gradient‑boosted classifier to predict cancellations.

A balanced view

Not every AI‑sounding initiative is empty. Some firms use the label as a catalyst to invest in genuine data‑science capabilities, gradually moving from static scripts to adaptive models. The key differentiator is transparency:

• Clear scope – Companies that specify whether a feature is rule‑based, statistical, or deep‑learning driven avoid confusion.
• Performance metrics – Publishing accuracy, latency, and error rates lets stakeholders assess the real impact.
• Ethical disclosure – Explaining how decisions are made, especially when they affect customers, builds trust and aligns with upcoming regulations.

Looking ahead

If the AI washing trend continues unchecked, the market may experience a backlash similar to the “green‑washing” episode in sustainability. Conversely, a shift toward honest communication could encourage a healthier ecosystem where genuine AI projects receive the attention—and funding—they deserve.

For practitioners, the practical takeaway is simple: use the term AI only when the technology truly learns from data. When in doubt, describe the solution in concrete terms—rule‑engine, RPA, statistical model—so that expectations match reality.

---

Related reading:

• The UK’s AI Regulation White Paper (https://www.gov.uk/government/publications/ai-regulation-white-paper)
• A guide to responsible AI marketing from the British Standards Institution (https://www.bsigroup.com/en-GB/AI-Responsible-Marketing)