Tanya Slyvkin highlights four AI brand risks for companies
Two lawyers faced a record $110,000 fine in 2026 after filing documents with AI-invented citations, an increase from the $5,000 in the 2023 case, with over 2,000 cases recorded worldwide per the ABA Journal. Templafy data shows 95% of knowledge workers edit AI documents before use. This context precedes the Whitepage announcement on four AI brand risks outlined by CEO Tanya Slyvkin.
Around 59% of UK and US workers in a Templafy survey indicate that AI-generated documents cause different businesses to sound overly alike, while just 30% note their tools automatically apply approved templates or brand assets. BCG research reveals 54% of employees would employ AI tools without company authorization. A Nature Human Behaviour study of 1.5 million Change.org petitions demonstrated that an AI writing tool rendered petitions more uniform without boosting success rates. Tanya Slyvkin, CEO and Co-Founder of Whitepage Studio, advises designating one person as the owner of truth to supply AI with brand tone, templates, and examples. She states, “AI tools know nothing about a brand unless someone gives it to them, so they tend to fall back on the same generic wording as everyone else.”
“Workslop” refers to AI outputs that appear polished yet lack depth. Stanford research indicates 40% of US desk workers encountered it recently, requiring nearly two hours to repair each instance at roughly $186 per employee monthly. Approximately half of recipients subsequently perceived the sender as less competent and dependable. A Science study showed 11 leading AI models provided affirmations 50% more frequently than humans, accounting for convincing but unchallenging drafts. Slyvkin recommends building drafts in stages on a single design system: narrative first, then rough prototype, then final design. She notes, “AI drafts tend to come back crowded with text and generic layouts, and that’s where the hours of fixing go.”
AI systems can generate plausible but false figures, quotes, and references. Findings from KPMG and the University of Melbourne show 66% of employees skip verification of AI output, leading 56% to commit errors. Deloitte encountered issues in two instances: an A$440,000 Australian government report in 2025 with AI mistakes and a CA$1.6 million report for Newfoundland and Labrador containing at least four fabricated citations. Slyvkin emphasizes verification, saying, “I really don’t like anything going to a client that nobody has traced back to a source. Before a document goes out, I’d have one named person check every figure, quote and reference, and if nobody can find where a number came from, it comes out.”
Mass AI content risks diminishing trust. A Seismic survey of 200 senior B2B leaders found 75% prefer human guidance over conflicting AI content. Research in Organizational Behavior and Human Decision Processes across 13 experiments revealed lower trust in individuals disclosing AI use. Slyvkin suggests marketing maintain a master deck with flexible sections for customer types. She comments, “Buyers can usually tell whether a deck was made for them or for everyone. Content built around one buyer’s situation reads as someone paying attention, while generic content tends to read as automated, even when a person wrote it.”
