Blog

Sumitomo Rubber Industries filed trademark application no. 586030 on 23 March 2023 for a floral fragrance reminiscent of roses applied to tires (Class 12). After examination objections under Sections 9(1)(a) and 2(zb), and with Shri Pravin Anand as amicus curiae, the CGPDTM (Prof. Dr. Unnat P. Pandit) allowed the mark for advertisement in the Trademarks Journal—India’s first accepted olfactory mark.

The Registry held that the scent was distinctive and that a 7-dimensional vector graphical representation (floral, fruity, woody, nutty, pungent, sweet, minty)—developed at IIIT Allahabad—was clear, precise, self-contained, intelligible, and objective under Section 2(1)(zb). The rose scent was arbitrary relative to tires, contrasting with the smell of rubber consumers expect on roads.

The order is registry-level (not appellate court) and draws on EU Sieckmann criteria; it opens a path for non-conventional marks in India while highlighting the need for formal guidelines on scope, examination, and enforcement as scents fade over time.

Generative AI can produce a brand identity in minutes, but trademark law does not treat machine output as a safe harbor. LLMs and diffusion models cluster on common words and visual motifs—so AI-generated logos and names can still confuse customers and infringe existing marks, whether or not the founder “intended” to copy anyone.

Skipping clearance searches before launch can amount to willful blindness. In first-to-use jurisdictions such as India and the US, an earlier unregistered user may force a costly rebrand even when your mark came from a prompt. The familiar Lush v Amazon line of cases shows how platform search and keyword behavior already tested “confusion”—now AI-mediated shopping shifts the question from the human eye to algorithmic gatekeeping.

Registries may reject marks not because AI made them, but because repetitive AI-driven designs fail distinctiveness. The article’s takeaway: use AI for inspiration; rely on human clearance, verification, and execution to protect the brand.

South Korea’s National Intelligence Service cited roughly ₩23 trillion (~$17 billion) in potential losses from attempted technology leaks in semiconductors. Such figures are policy signals—gross macroeconomic impact—not defensible valuations; they rarely reflect enforceability, litigation outcomes, or the half-life of technology advantage.

A plain 5-year DCF at 10% WACC on evenly distributed $3.4B annual flows yields about $13B present value. Applying IP-specific multipliers (enforceability 0.6, market control 0.8, advantage life 0.9, litigation environment 0.7) reduces realistic recoverable or monetizable value to roughly $4B.

Neither number is “wrong”—they answer different questions. Headline estimates mobilize attention and budgets; adjusted valuation supports investors, courts, and licensing strategy. Understanding true value at risk requires qualitative IP variables, not sensational totals alone.