Trademark Search, Artificial Intelligence and the Role of the Private Sector

Intellectual Property, Copyright and Trademark and Artificial Intelligence

Article Snapshot

Author(s)

Sonia Katyal and Aniket Kesari

Source

Berkeley Technology Law Journal, Vol. 35, pp. 501-586, 2021

Summary

Worldwide, trademark offices and private firms use artificial intelligence-based systems (AI) to identify distinct trademarks. AI will transform trademark business and legal processes.

Policy Relevance

AI allows firms to calculate the litigation risk of different trademarks.

Main Points

  • AI can help businesses select a distinctive trademark; registration of a trademark that is confusingly similar to an existing mark is prohibited by the Trademark Act.
     
  • Trademark offices worldwide increasingly use AI to identify similar marks; AI could replace some human analysts, but a mix of human and non-human analysis will produce the best outcomes.
     
  • Observers expect AI to streamline the trademark application process, improving trademark quality, but making marks harder to obtain; AI could detect trademark infringement or design subtle counterfeits to deceive consumers.
     
  • The government-supported Trademark Electronic Search System (TESS) is incomplete, and private-sector trademark search engines have emerged to complement TESS.
     
  • A series of experiments tested the efficacy of the private-sector search engines using marks rejected by the Patent and Trademark Office (USPTO) because of similarity to existing marks.
     
    • The private systems returned fewer results than TESS, filtering out irrelevant results.
       
    • They performed as well as the USPTO at finding at least one conflicting mark.
       
    • They performed better than the USPTO at finding all conflicting marks.
       
  • AI-based search engines can predict whether a trademark can be successfully registered or is likely to result in litigation, but cannot assess risks such as tarnishing or dilution of trademarks.
     
  • Trademark registration should be treated as an adversarial machine learning problem because, over time, AI search tools and the USPTO will continually adapt to one another’s decisions.
     
  • Traditionally, trademark analysis focusses on economic factors such as the role of trademark in reducing the cost to consumers of finding products, and the effects of trademark on demand.
     
  • Trademark applicants are also consumers of trademarks.
     
  • Overly broad trademark protection can stifle competitors merely seeking to describe their own products.
     

Get The Article

Find the full article online

Search for Full Article

Share