Very few businesses operating today do not use some kind of AI in most aspects of their operations. Therefore, the question of how to use AI to protect intellectual property, especially in global supply chains, requires some definition.
The term "artificial intelligence" is broad and includes everything from Alan Turing's 1947 "machine that can learn from experience" to today's GPT models that generate their own content based on human prompts. Here, AI means the use of natural language processing and large language models like the GPTs to search for and recognize possible infringement.
A critical step to prevent harm from IP theft is to identify infringing goods or services. With its ability to process vast amounts of data and recognize patterns, AI can proactively identify potential IP infringements in ways that human monitoring cannot match.
Some examples of proactive monitoring can include:
- Using AI algorithms to analyze online data to identify possible infringements. This can include counterfeit goods being sold on e-commerce platforms, or pirated content on digital channels.
- Using Natural Language Processing to scan and interpret text across various online platforms, flagging instances where copyrighted or trademarked content is used without authorization. This NLP text scanning can also be used to find instances where trade secret materials are being sent—for example, to an employee's own address or to a competitor's email to permit copying.
- Using AI to recognize images and videos, making it easier to spot unauthorized use of visual content.
DEFEATING DEEP FAKES
A rise of deep fake technology to mimic images and voices results in understandable human error when sending sensitive information to competitors or thieves. AI tools can detect whether messages requesting sensitive information come from a generative AI source instead of a human counterpart.
In addition to these models, I advocate for an older but still effective technology involving smart contracts, blockchain or other distributed networks and, increasingly, NFTs. The proliferation of advanced manufacturing relies on transmitting data to manufacturing facilities around the globe. Smart contracts are if/then software programs that are secured to a blockchain to verify that they have not been modified or that the data wrapped in the smart contract has not been copied. Where the IP is encoded into digital manufacturing specifications, the IP cannot be taken off the blockchain without the theft being evident to the IP owner.
NFTs' more interesting capability is moving data in a way that is highly traceable. Newer versions of NFTs, called vNFTs or virtual NFTs, can be used to verify identity of trading partners in transactions, and wrap highly sensitive data in an immutable digital package for secure transmission and use.
Finally, identifying IP theft is not enough to function as a robust deterrent and remediation program. Enforcement is a necessary component of any IP protection program and in this aspect both AI and blockchain-tethered smart contracts serve a valuable function.
By using computer algorithms and immutable data provenance mechanisms to detect AI theft, especially taking trade secrets via some form of electronic data transfer, lawyers can more easily track and prove how data was transferred. In addition, lawyers regularly use AI to assist with research into infringing methods and apparatus so that proving infringement and defending against infringement claims has become more efficient and effective.
IP protection is a fascinating problem with equally interesting solutions.
Originally published 28 September 2024
The content of this article is intended to provide a general guide to the subject matter. Specialist advice should be sought about your specific circumstances.