Thailand, renowned for its rich cultural heritage, has always embraced artistic endeavors and innovative thinking. With the advent of AI, a new wave of creative possibilities has emerged, pushing boundaries and challenging traditional notions of creativity. AI-generated works harness the power of machine learning algorithms to analyze vast amounts of data, enabling computers to mimic human-like creative processes and generate unique outputs.
One significant challenge is when companies seek to protect their AI technologies and innovations. Some companies have sought to patent GPT-related technologies. For example, OpenAI, the company behind GPT, has filed several patents related to GPT, including patents for natural language processing systems and methods, as well as patents for language models and language generation systems. Other companies have also filed patents for AI-related technologies, including Google, IBM, and Microsoft.
In this article, we will explore five key steps to protect AI-related technologies in Thailand, helping businesses secure their valuable intellectual assets.
Step 1: Conducting a Prior Art Search
Before diving into AI development, it is essential to conduct a thorough prior art search. This search aims to identify existing patents, publications, and other relevant information that may impact the novelty and inventiveness of your AI technology. By identifying prior art, you can assess the potential patentability of your invention and avoid infringement issues.
Step 2: File for Patents
The patentability of AI-related technology requires careful consideration of the technical features of the invention and its effects on the physical world. The criteria for patentability are designed to ensure that the technologies that are patentable provide clear technical contributions to the invention. By understanding these criteria, inventors and companies can increase the chances of successfully obtaining a patent for their computer-related inventions.
One of the most effective ways to protect AI-related technologies is through patents. Nonetheless, the main challenge of patenting AI systems is the issue of inventiveness. The patent system requires that an invention be novel, non-obvious, and useful. However, with AI systems, it is difficult to determine what constitutes inventiveness. Since most AI-generated work is pre-trained on existing data, it may not be possible to determine whether a particular output is truly novel or simply a recombination of existing information.
Another significant challenge when it comes to patenting AI-generated work is the issue of disclosure. Patent law requires that an invention be fully disclosed in the patent application. However, with AI-generated work, it may be difficult to provide a full and complete description of the invention. This is because AI systems like GPT are based on complex algorithms and deep learning models that are difficult to explain in layman's terms. Furthermore, disclosure of the algorithm itself may expose the system to reverse engineering, undermining the novelty of the invention and exclusivity to invention which the patent grants.
In Thailand, under Section 9(3) of the Thai Patent Act, software and computer programs are not patentable. However, there are limited circumstances wherein a computer program may be patentable. These criteria are as follows:
- The software and hardware work in tight collaboration in their technical aspects.
- A program that provides technical contributions to the invention.
- A computer that is not used as only a medium for processing or storing data.
- An output/calculation of the program that is not for business-operation practices.
These criteria are designed to ensure that computer programs that are patentable provide technical contributions to the invention and have clear technical effects that are linked to an identifiable physical nature.
For example, the first circumstance is a computer program that is designed to work with a specific device, machine, or process, but not solely with a normal, non-specific computer. One example would be a computer program for a process control device/valve.
The second circumstance is a computer program that technically improves hardware. The program would need to have a key role in improving the efficiency of the hardware. One example would be a computer program that controls an ABS braking system.
According to the Thai Patent Law, a computer-related invention may be patentable if the patent application claims an invention that is not a computer program per se, such as source code or object codes, or a mere computer-readable storage containing a computer program. Instead, the invention must have technical features that can be established by showing that the claimed invention includes hardware or physical mechanisms, is operably connected to hardware or physical mechanisms, and/or produces clear technical effects that have a direct link to an identifiable physical nature.
Step 3: Copyright Protection
Copyright can protect original works of authorship, including software programs used in AI systems. In Thailand, copyright protection is automatic upon creation, but registering your copyright with the Copyright Office can provide additional benefits. By registering your AI-related software programs, you gain stronger evidence of ownership and the ability to enforce your rights against potential infringers.
Step 4: Protect Trade Secrets
While patents provide strong protection, some AI-related technologies may qualify as trade secrets. To safeguard trade secrets, companies should implement robust confidentiality measures. This includes implementing access controls, non-disclosure agreements (NDAs), and secure storage systems to prevent unauthorized access and disclosure of valuable AI algorithms, datasets, or proprietary techniques.
Step 5: Collaborate and Establish Partnerships
Collaborating with other organizations can foster innovation in AI technologies. However, it is important to ensure that proper legal agreements, such as technology transfer agreements or joint development agreements, are in place to protect your IP rights. Clearly, any agreement with third parties must define ownership, confidentiality, and licensing terms in such agreements to prevent disputes and unauthorized use of your AI technologies.
Below are the top five companies with AI-powered tools in Thailand.
- AppMan Co., Ltd.: A leading AI and data science company specializing in developing AI-powered solutions and platforms for various industries such as finance, healthcare, and e-commerce.
- Zensorium Co., Ltd.: A Thai company that focuses on computer vision and AI technologies, providing solutions for object recognition, video analytics, and intelligent surveillance systems.
- Builk One Group: An AI-driven construction technology company, offering innovative solutions for construction project management, leveraging AI, machine learning, and data analytics.
- Omise Holdings: A fintech company that incorporates AI into their payment gateway solutions.
- AiServe Technologies Co., Ltd.: A Thai startup that specializes in developing AI-powered robotics and assistive technologies for the visually impaired. Their solutions include AI-based navigation systems and smart mobility devices.
Protecting AI-related technologies in Thailand requires a proactive approach to intellectual property. By conducting prior art searches, filing for patents, safeguarding trade secrets, leveraging copyright protection, monitoring the marketplace, and staying updated with IP laws, businesses can secure their valuable AI innovations. Engaging IP professionals and legal counsel with expertise in AI and IP can provide invaluable guidance throughout the process. Ultimately, safeguarding intellectual property rights encourages innovation, fosters economic growth, and positions businesses for success in Thailand's AI landscape.