Listen to this articleMany industries worldwide experience intensive competition because market leaders and potential disruptors are used to use artificial intelligence in products, services and business processes. Driven by considerable investments and fast innovations, AI tools are already used to automate high-ranking tasks, increase productivity and to manage the decision-making in industries that range from health care to transport to the construction area.
These investments in software and data science naturally raise questions about the protection of AI assets. Getting intellectual property for AI assets can help protect the market position or determine partnership conditions. IP can help to raise interest from investors, to fulfill freedom, to fulfill the claims during the duty of care, to define the scope of strategic license contracts, to prevent competitors from determining/expanding their market committees, or a reaction to the IP intersperson claims of a competitor and more. If a company is created or used, it should evaluate how IP supports these goals.
While the protection of other technologies can be relatively simple, the protection of AI work can be complex in some cases. In view of the technical and legal complexities, the defense company and valuable IP protection are based on experienced lawyers who adapt the IP strategy through a deep understanding of the product and business.
For companies that use IP to promote business interests instead of receiving patents as trophies, a detailed discussion with the advice of ideas, the competitive market, the business goals and the product design of the keys to defending and estimated protection for AI.
Defense ability, value of AI IP
For companies with well-established and robust processes for the identification and securing of IP rights in valuable innovations, adjustment or fine-tuning of these processes may be necessary in order to reliably obtain strong IP rights in AI assets.
In the United States, the legal landscape in relation to the acquisition and enforcement of IP rights in AI innovations is already complex and continues to develop quickly. There is therefore nuanced and AI-specific considerations.
Despite the latest uncertainty about the patentability of some types of software, the AI software and workflows have patented both in Europe and in the US leading judicial decisions and administrative guidelines.
In view of the technical and legal complexities, the defense company and valuable IP protection are based on experienced lawyers who adapt the IP strategy through a deep understanding of the product and business.
In many industries, business secrets are also an increasingly effective option for the protection of AI assets -although the disclosure requirements in heavily regulated industries are the opportunity to rely on business secrets to protect some AI assets.
Nowhere is the expression “Data is the new currency” more true than with high -quality, anonymized data sets that collect and use companies in product development. Especially if a product uses conventional hardware or executed on the hardware of another, it can be of crucial importance for a company to obtain IP for AI, software and data.
AI IP is also valuable. In October 2023, a smaller healthcare company received an order from the US International Commerce Commission, which blocked the import of the smart watches of a large computer company in the validity and violation of the AI-related patents in the court hearing. This larger company ultimately has health monitoring technology from the order from its watches, which restricted the competition for the smaller company.
One of the greatest business secrets that concentrated on AI technology for the use of autonomous vehicles in which a company has agreed to pay hundreds of millions of dollars after the allegations that a new employee had paid a former employer was reduced by a former employer.
With a large variety of AI data that has been used by the proper licensed data that has been used for the training -KI models, many companies try to licensed their data to model developers, and model developers strive to licensed data records.
Avoid pitfalls and secure high -quality IP
Against this background of the high value and the concrete protection, companies can still face challenges.
Some AI systems risk calling the unpatentable automation of existing manual processes. In view of the fact that many of the basic AI techniques that have been used in today's state-of-the-art products (including “deep” learning), years or decades old, previous work can occur.
The detection of violations can be difficult for back-end functionality. In some cases, certain AI techniques are replaced by short duration and easily by improved or alternative versions.
Instead of preventing companies from protecting their AI assets, these challenges should influence the development of an IP strategy that gives added value and avoids waste. Experienced IP consultants with specialist knowledge in AI can analyze these factors together with product details and business goals, conduct in-depth discussions that are necessary to superior to patents with value or to supplement patents with other IP, and to build a tailor-made IP strategy that helps the company achieve its goals.
IP, which offers the best protection for AI assets, may often not be on the AI itself. The first thoughts of the companies are often a specific model that they have designed or trained, but the exact model can be relatively short -lived. The model is retrained and in some cases the structure of the model is improved. While your team's selected model does well, alternatives can be available and determine whether a competitor uses your or another model. Business secrets can be appropriate for certain high-quality models, and the patent strategy can concentrate more on other aspects of the product workflow that interact with AI.
AI outputs usually drive through the downstream processing and lead to diagnostic expenditure (e.g. in healthcare and cyber security), to implementable predictions (e.g. in marketing and retail), control signals for systems with closed loops (e.g. in manufacturing, transport, logistics, home automation and energy supply), recommendations for users, etc.
Although the models develop quickly, AI-controlled workflows can exist across product ceremonies. In contrast to models, these workflows are often clearly visible to customers, are usually the focus of marketing campaigns and the buying decisions of customers and can be recognized by copies of competitors – all factors that increase the patent value.
Data rights and licensing for these AI outputs also offer high value, especially in the context with corresponding inputs. The data preparation processes that create the entries for AI models, such as: B. Data treatment and preliminary processing of raw data in informative features can also be fertile IP sources. These contexts for input output can offer important support for patent protection.
It can also be possible to take into account commercial protection in which features or processing are not visible. The data rights for access to the input data can be invaluable, especially for training data received by partners. Such a non-patent IP can be particularly important if the AI largely imitates an existing manual process.
Next Steps
The development of an effective IP strategy for AI assets should include detailed discussions with IP consultations on how the company uses AI internally and customer-oriented products and services to advance the business goals.
Such cooperation is the key to determining which IP species offer the most defensible and valuable protection for each AI assets.
Andrew “AJ” Tibbetts is the shareholder in the group for intellectual property and technology in the Greenberg office sad in Boston. He is a former software engineer. It can be contacted at [email protected]. Samuel S. Stone, Associate in the Boston Office of Greenberg, is a member of the risk capital and the aspiring technology practice and innovation and the artificial intelligence group. It can be contacted at [email protected].