ChatGPT is an AI-based language model that can generate human-like responses to text prompts. It can be used in a variety of ways, including customer service, marketing, and HR. In HR, ChatGPT can be used to streamline various tasks, including recruitment, employee onboarding, and training.

Applications

Recruitment is one of the most critical HR functions, and ChatGPT can be used to make the process more efficient. It can help HR departments via:

• creating professional job descriptions based on a set of criteria and company templates

• comparing CVs to job descriptions and providing a percentage match between the candidate CV and the job description,

• respond to simple candidates’ queries about the onboarding process,

• and more.

At the next step, employee onboarding, ChatGPT can help new hires get up to speed quickly by providing them with access to training materials and answering any questions they may have. To answer organization specific questions, AI-based language models need to be trained on the organization specific data as, out of the box, non-public company information will not be included in the AI model’s training data.

In Saudi Arabia and the GCC, the above applications are being explored carefully. For example, companies typically have no issue using ChatGPT to create job descriptions as this information typically does not have any privacy concerns associated with it when used with ChatGPT, it will be made publicly available when the position goes live. However, submitting a candidate’s CV to ChatGPT for comparison with job descriptions is a concern as it contains a person’s private information.

Although companies are in the early stages of reviewing ChatGPT applications, across Saudi Arabia and the GCC, organizations have already begun exploring other non-ChatGPT usages of AI such as carrying out sourcing through recruitment agency marketplaces which leverage artificial intelligence to match job requirements with an agency’s specialization to find the best candidates. These applications are a simple entry point for many GCC companies to begin using AI.

Intellectual Property Concerns

When using ChatGPT, it is essential to consider intellectual property (IP) considerations. If the AI system is being used to generate original content, such as training materials, it is essential to ensure that the organization has the proper rights (as described below) to use that content and also that the organization is aware that any data submitted to ChatGPT in this process can be included in its training data, which has led to data breaches at some large enterprises.

In terms of the generated content such as training materials, the legal rights to use ChatGPT’s content fall under the OpenAI API Terms of Service. According to these terms, users are granted a limited, non-exclusive, non-transferable, non-sublicensable license to access and use the API and any content generated by the API solely for their internal business purposes.

In Saudi Arabia, organizations can further protect their IP by registering their copyrights and trademarks with the Saudi Arabian Ministry of Commerce and Investment, however, care needs to be taken especially with AI generated content. As IP can be generated so quickly with tools like ChatGPT, and countries like Saudi Arabia follow a first-to-file IP system, this means organizations need to register IP first to have legal rights over that intellectual property. Therefore, it’s important for organizations to register their IP as early as possible to avoid potential conflicts and legal challenges in the future, especially inventions core to their business.

In the last months we have seen the attempts by individuals to register inventions by AI as owned by the AI system itself. For example, a computer scientist Stephen Thaler created an AI system called DABUS to create new inventions (e.g. beverage holder and emergency light beacon) and then applied for patents on these inventions under DABUS’ name. This has led to patent rejections in a number of countries such as:

• Australia

• UK (still under appeal at the time of writing)

• New Zealand

• USA (rejected by US Supreme Court on 24th of April)

In Saudi Arabia, however, a decision has not been made as to whether the AI system DABUS can be the inventor. Hence, caution needs to be taken when dealing with AI generated intellectual property.

Ownership and Liabilities

Regarding authorship of materials produced using ChatGPT in Saudi Arabia, the answer will depend on the specific circumstances. Most organizations in Saudi Arabia have moral and legal rights to all works created by an employee as part of their employment. If this is not the case, then in some cases, the employee who inputs the prompts may be considered the author.

For example, if an employee of a company uses ChatGPT to generate content for a personal project, they may be considered the author of the resulting work, and they would have the copyright ownership of that content. However, if the employee is using ChatGPT to generate content as part of their job responsibilities or if the content is created using company resources, the organization may be considered the author of the work and own the copyright to the resulting content.

If inaccuracies are present in the material produced by ChatGPT or similar AI products in Saudi Arabia, the potential liability will depend on who is responsible for the inaccuracies. If the inaccuracies are a result of the AI system itself, the liability may lie with the organization that developed the system. However, if the inaccuracies are a result of employee input, the liability may lie with the employee who entered the incorrect information. This also is an area of flux.

The New Future

ChatGPT and similar AI products are revolutionizing the way HR departments operate in Saudi Arabia and the GCC. However, it is essential to consider IP considerations and ownership of content when using these systems. Additionally, organizations must have clear guidelines in place regarding liability for inaccuracies to avoid any legal issues and ensure they protect their data during the usage and training of such systems.

Written by Dr Evan Shellshear.