Key concepts in AI: An expert glossary for IP professionals

Artificial intelligence (AI) is no longer a futuristic concept; it is a present reality already positively impacting intellectual property (IP) practice. AI technologies enhance the creation, maintenance and protection of IP assets by providing tools that streamline processes, improve accuracy and supercharge decision-making.

Despite its potential, AI and its associated concepts are not always familiar to a profession that is more used to dealing with creating, managing and protecting intellectual property (IP). Day-to-day AI can feel like a nebulous and inaccessible opportunity.

To help patent and trademark professionals understand AI and its associated technologies in context, we've compiled a concise, alphabetical glossary of key terms with relevant examples that bridge the gap between the promise of AI and the practicalities of its use in practice.

Key concepts in AI

1. Artificial intelligence

AI is a broad field that includes various technologies, such as machine learning, deep learning, natural language processing and computer vision.

AI in IP: Since AI is an umbrella term for many enabling technologies, it's helpful to consider the impact of AI in the context of a set of tasks. A good example is the trademark examination workflow. As shown in the below visual, AI can assist a trademark professional throughout the examination workflow. This includes:

• Searching (text and image),
• Classification (assigning the correct codes),
• Rules-based checks (absolute grounds and jurisdiction),
• Examination (word marks, image applications, and past precedents),
• Appeals (risk-scoring based on previous Office decisions)
• Policy (global and regional trend analysis).

AI enhances these functions without adding extra steps.

Image: A visual representation of where AI can impact tasks in a trademark examination workflow

2. Automation

Automation uses technology to perform repetitive tasks typically done by people.

Automation in IP: Automation tools in patent drafting solutions manage routine tasks like matching PTO records and entering information into standard forms.

3. Deep Learning

Deep learning (DL) uses neural networks (NNs) with many layers (hence 'deep' networks) to analyze data and identify intricate patterns simpler algorithms might miss.

DL in IP: DL models can efficiently analyze complex patent documents, identify key concepts and categorize large patent portfolios.

4. Generative AI

Generative AI (GenAI) models 'generate' new content based on training data.

GenAI in IP: GenAI can create content for patent drafting, search for trademarks and prior art, review promotional materials, train by generating insightful answers and strategize through analysis and suggestion.

As with any task that requires the creation, analysis and review of expert content, human oversight and review must accompany the use of GenAI tools. This oversight is necessary to mitigate known risks (like hallucination and bias), maintain control (both in output and input) and ensure the foundation of the content is correct and precise.

5. Large Language Models

Large language models (LLMs) understand and generate human-like text based on large datasets.

LLMs in IP: LLMs can assist in drafting patent applications by generating text and ensuring consistency.

6. Machine Learning

Machine learning (ML) uses algorithms and statistical models to help computers improve their task performance through experience.

ML in IP: ML algorithms can analyze trademark data to identify trends and assess risks, speeding up decision-making processes.

7. Neural Networks

NNs are computing systems inspired by the human brain and excel at handling complex tasks. Convolutional neural networks (CNNs) are a subset of NNs that use convolutional layers that act like filters to detect patterns in images and other data, much like how our eyes process visual information.

NN in IP: AI-powered trademark research solutions use NNs to help evaluate the risk of conflicts by assessing an owner's litigation behavior and portfolio reach.

8. Natural Language Processing

Natural language processing (NLP) focuses on the interaction between computers and humans through 'natural language.' It's about teaching computers to understand, interpret and respond to text or speech in a way that feels natural to humans.

NLP in IP: Chatbot-style AI tools frequently use NLP to answer common natural language questions.

9. Transformer models

Transformer models are deep learning models that revolutionized NLP by using attention mechanisms (techniques that allow the model to weigh the importance of different parts of the input data) to process sequential data. Unlike traditional models, transformers can focus on different parts of the input data simultaneously, understanding context and relationships more effectively.

Transformer models in IP: Many LLMs rely on transformer models for tasks like automated patent classification and semantic search.

10. Responsible AI

Responsible AI refers to the ethical, transparent and accountable development and use of AI systems. It focuses on several key principles:

• Fairness: Ensuring AI systems do not perpetuate or amplify biases.
• Bias mitigation: Actively identifying and reducing biases in AI models.
• Explainability: Making AI decisions understandable to humans.
• Transparency: Providing clear insights into how AI systems operate.
• Continuous monitoring: Regularly checking AI systems for performance and ethical compliance.
• Regulatory compliance: Adhering to laws and guidelines governing AI use.

Responsible AI involves collaboration with experts like data scientists, IP practitioners, risk and compliance managers and technologists who can access clean and curated data to ensure accuracy and minimize risks.

Responsible AI in IP: Since responsible AI is an approach, it should frame the development and deployment of AI-enhanced IP tools to avoid biases. It also helps ensure these tools are explainable and promote transparency and fairness across all industries and technologies.

A practical approach to AI

This glossary is designed to help IP professionals quickly grasp AI-related terms and enhance their understanding of how these technologies can be applied to daily tasks.

For a deeper review of AI's practical applications in IP, we invite you to read our latest white paper, "Artificial intelligence for the IP legal profession: Practical approaches for harnessing the potential of AI." This comprehensive guide explores real-world use cases, benefits, and considerations for integrating AI into IP workflows.

Download your copy today: Future-proof your IP practice with AI

Public link

Please use the above public link if you want to share this noodl on another website.