The rise of Artificial Intelligence (AI) has been largely driven by one tool in AI called machine learning. In 1959, Arthur Samuel defined machine learning as the “field of study that gives computers the ability to learn without being explicitly programmed”. In the context of artificial intelligence, machine learning is a software that enables computers to learn from a table of data (dataset) and improve its performance over time without explicit programming.
The most common type of machine learning used is a type of AI that learns A to B, or input to output mapping. An example of this type of machine learning is when an AI program generates a text transcript (output) from an audio clip (input). While this type of supervised learning has been around for some time, the open-source nature of the internet and computers allow for the unprecedented gathering of enormous amounts of data. The data is the integral piece that propels the rapid advancement in AI and other large language models.
Licensing Issues and Risk Mitigation
There are steps you can take as a business owner or principal to mitigate risk while ethically integrating AI and other large language models into your business. 2024 will be the year we carefully watch the courts undertake the very early stages of pending copyright cases. As mentioned on Reuters, U.S. Supreme Court’s Chief Justice John Roberts emphasized that “any use of AI requires caution and humility.” At the end of the day, legislative and regulatory change will likely be the key components to put up guardrails for AI.
In the midst of uncertainty as the courts attempt to catch up with the implications of AI technology, there are certainly measures you can take and best practices to consider when using AI learning for your business.
The risk of using non-commercial licensed materials for a commercial purpose is at the heart of the issue when using data gathered by open internet to be fed to AI models. Here are things to consider when using existing data for AI development. Ideally, you will want to consider and have the opportunity to obtain clarifications to the following issues:
- Did the licensor properly obtain the data?
- Does the licensor have sufficient rights to the underlying data?
- Does the licensing agreement provide appropriate non-infringement representations and warranties?
- And if the licensing agreement does provide appropriate non-infringement representations and warranties, whether the licensor will indemnify the licensee (your business) if infringement issues arise from your use of the data.
Before making important business decisions involving AI gathered and generated information or products you may want to reduce your risk by consulting an experienced business and trademark attorney for clarification and guidance. The attorneys at Quadros, Migl & Crosby are here to help when you have questions. Please call (713) 300-9662 to schedule your consultation.