The UKIPO launches AI-powered assessments of trademark applications

The UK Intellectual Property Office (UKIPO) is testing an AI-driven tool to help the public apply for trademarks. The UKIPO’s tool is intended to be similar to Australia’s Trade Mark Assist tool and the UKIPO worked with the Australian IP Office to accelerate development.

According to an UKIPO post (How can Artificial Intelligence improve customer experience at the IPO?), this new solution will include:

• an interactive classification picker to help applicants identify the class of goods and services and to calculate application costs;
• AI-enabled checks indicating objections that might be raised against their proposed trademark; and
• an AI-based search for similar trademarks that may prevent registration.

The World Intellectual Property Organisation (WIPO), the European Patent Office (EPO) and many national Intellectual Property offices are studying the potential of AI to improve their processes. It was reported in 2017 that the Japanese Patent Office (JPO) reviewed over 800 aspects of its procedures to identify those for which AI presented the greatest cost-benefit. WIPO has an ongoing review of AI in “Intellectual Property Administration” and has published an index of 64 initiatives from around the world. It is exploring:

• Automatic classification of patents and goods/services for trademark applications;
• Search of patent prior art and figurative elements of trademarks;
• Examination and formalities checks for trademarks and patents;
• Helpdesk services (automatic replies to client); and
• Machine translation, linguistic tools and terminology.

Such advances are not always straightforward: at the 2019 AIPPI symposium on Artificial Intelligence, it was noted that the United States Patent and Trademark Office has over 8,000 examiners, so even system stability is a challenge and such IT challenges are more intense for processor-intensive AI applications. If the promise of AI is achieved, applicants may see faster grants of patents and trademarks, but potentially more problematic prior art and earlier marks found during examination. We may even see an arms race between IP offices and sophisticated applicants. There are persistent rumours that some technology companies have applied machine learning to prosecution records to identify patterns in the work of individual examiners, such as their favoured objections and the counter-arguments that have been most successful with them in the past.