AI-Generated Writing Exposed: The Alarming Surge of Synthetic Patterns in Corporate Communications

SAN FRANCISCO, CA — April 30, 2025: A distinctive linguistic fingerprint is now exposing the pervasive use of artificial intelligence in corporate America. The sentence construction “It’s not just X—it’s Y” has transformed from a common rhetorical device into a statistical marker for AI-generated content, according to recent market analysis. This pattern’s dramatic proliferation in official business documents signals a fundamental shift in how companies communicate with investors, regulators, and the public.

The Statistical Evidence of AI-Generated Writing

Market intelligence firm AlphaSense conducted a comprehensive scan of its extensive database, which aggregates corporate news releases, earnings reports, and regulatory filings. The findings, initially reported by Barron’s, reveal a startling trend. Instances of the “not just—but also” construction surged from approximately 50 mentions across major corporate documents in 2023 to over 200 uses by early 2025. This represents a more than fourfold increase in just two years.

This data provides a quantifiable metric for a phenomenon many linguists and communications experts have observed anecdotally. The construction serves as a linguistic tic common in large language model outputs, often used to add rhetorical flourish or emphasize a dual benefit. Consequently, its sudden frequency acts as a proxy for measuring AI’s penetration into formal business writing.

Real-World Examples from Major Corporations

The trend is not confined to obscure documents. Analysis shows its adoption by leadership and communications teams at some of the world’s most prominent technology and consulting firms. These examples illustrate the pattern’s widespread use:

• Cisco stated, “In 2025, AI won’t just be a tool; it will be a collaborator.”
• Accenture proclaimed, “The future of autonomy isn’t just on the horizon; it’s already unfolding.”
• Workday noted, “DevOps teams are managing not just deployments, but also security compliance and cloud spending.”
• McKinsey & Company reported, “These systems aren’t just executing tasks; they’re starting to learn, adapt, and collaborate.”

Microsoft’s corporate blog featured multiple instances in a single post. CEO Satya Nadella was quoted, and the blog’s narrative heavily relied on the structure, highlighting its utility in framing technological ambition.

Linguistic Analysis and Origin of the Pattern

Experts in computational linguistics explain this phenomenon as a reflection of training data. Generative AI models learn from vast corpora of human-written text, which includes this effective but now overused persuasive structure. Dr. Elena Torres, a computational linguist at Stanford University, notes, “These models optimize for patterns that statistically predict ‘good’ or ‘complete’ sentences. The ‘not just—but also’ framework is a strong signal for adding informative contrast, making it a high-probability output for certain prompts.”

The pattern’s rise coincides with the corporate adoption of AI writing assistants for drafting press releases, investor summaries, and internal reports. These tools promise efficiency and polish but can homogenize language across industries. Furthermore, the em-dash (—), a key component of this construction, has itself become a secondary tell for synthetic text, as human writers increasingly favor shorter sentences and different punctuation in digital formats.

Broader Implications for Transparency and Trust

This trend extends beyond a simple curiosity about language. It raises significant questions about transparency, authenticity, and liability in corporate communications. Securities lawyers point out that documents like earnings reports and SEC filings carry legal weight. The undisclosed use of AI in their creation could potentially complicate issues of accountability and intent.

Communications professionals are now grappling with this new dynamic. “The goal is clear, compelling communication,” says Michael Chen, a partner at a global PR firm. “If AI helps achieve that, it’s a tool. However, over-reliance that leads to detectable, formulaic language can undermine the very authenticity and trust we aim to build. It’s a strategic risk.”

The table below summarizes the key shifts identified by analysts:

The Future of Human and Machine Collaboration

Looking ahead, the focus is shifting from detection to integration. The challenge for businesses is not to avoid AI tools but to use them strategically while preserving a human editorial voice. This involves training teams to prompt effectively, edit critically, and understand the limitations of generative models. The next phase may see the development of more sophisticated AI that can mimic a wider variety of human styles or corporate-branded voices, potentially making such simple detection methods obsolete.

Meanwhile, detection technology is also advancing. Startups and academic labs are building classifiers that look at hundreds of linguistic features—including syntax, word choice, and even semantic coherence—to identify synthetic text with greater accuracy than any single phrase ever could.

Conclusion

The epidemic of the “not just—it’s also” construction in corporate documents is more than a linguistic curiosity. It serves as a measurable indicator of the rapid, often opaque, integration of generative AI into the core channels of business communication. This trend highlights a critical juncture for corporate transparency, legal responsibility, and the enduring value of authentic human voice in an automated world. As detection methods evolve, so too must corporate policies regarding the use of AI in crafting the messages that shape market perceptions and stakeholder trust.

FAQs

Q1: What is the main evidence that AI is being used in corporate writing?
A1: The primary evidence is a quantitative analysis showing a more than 400% increase from 2023 to 2025 in the use of the specific sentence construction “It’s not just X—it’s Y” in official documents like earnings reports and news releases, a pattern highly correlated with AI text generation.

Q2: Why do AI models frequently generate this particular phrase?
A2: AI models are trained on vast amounts of human text. This construction is a common and effective rhetorical device in persuasive and explanatory writing, so the models learn it as a high-probability way to create contrast and add emphasis, leading to overuse.

Q3: Does using AI to help write corporate communications violate any regulations?
A3: Currently, there is no specific regulation prohibiting the use of AI assistance. However, undisclosed use in legally material documents like SEC filings could raise future questions about accountability, authorship, and the application of existing disclosure and liability laws.

Q4: Besides this phrase, what are other signs of AI-generated text?
A4: Other indicators include overuse of em-dashes, unusually consistent and formal tone, predictable paragraph structures, a lack of original metaphors, and sometimes factual inconsistencies or “hallucinations” within otherwise fluent text.

Q5: How can companies use AI for writing while maintaining authenticity?
A5: Companies can maintain authenticity by using AI as a drafting or editing tool only, ensuring strong human oversight and final editorial review. They should train staff to prompt for varied outputs and to inject company-specific voice, data, and nuanced perspective that AI may lack.