Responsible Usage of Generative AI: Professional Standards and Integrity

In the 59th edition of?this newsletter, entitled “Data Protection & Safety Measures of Generative AI: How to Maintain Data Privacy and Security While Using Public LLMs Services,” it was concluded that the primary concern about using the?already existing?“General-Purpose Large Language Models”?services is usually the?“Leakage of Sensitive Data”?to these models and the ability of these generative AI models to train themselves on these leaked data and learn from this data too. Although identifying sensitive data in each enterprise seems simple, from the point of view of all?“Data Protection Regulations,”?some non-sensitive information may become so sensitive, such as?“Personally-Identifying Information—PII,”?that should be dealt with so many restrictions. By mistake or due to lack of education, employees may upload it to the public LLMs application. Unfortunately, this totally wrong action can’t be prevented by any enterprise security measures. Only?“Well-Educated Employees”?can apply data protection and security measures.

In addition,?in the 60th edition of?this newsletter, entitled “Governance of AI Systems vs. Governance of Using the AI Systems: The Challenging Paradox,” it was concluded that the enterprise AI strategy should clearly distinguish between the?“Governance of AI Systems”?and the?“Governance of Using AI Systems.”?Although the?“Governance of AI Systems”?has a narrower scope and may appear as a straightforward strategy, the?“Governance of Using AI Systems,”?on the other hand, encompasses a broader range and more challenging considerations, such as ensuring that other non-enterprise AI systems are not being used irresponsibly. Enterprises need to manage this?“Challenging Paradox of AI Governance”?carefully.

Both of these previous editions of the newsletter tried to deal with professional standards that should be followed regarding the?“Input data”?when using the public general-purpose lange models services. However, this edition of the newsletter will focus on the professional standards that should be followed regarding the?“Output Data”?of the public general-purpose large models services as well as the?“Overall Process”?of using them.?

The most critical aspect regarding the output data is the domain of usage of this data. Suppose it will be used internally or externally by the enterprise. External use of this output data is usually subject to?“Intellectual Property and Copyrights Regulations.”?For example, If the public generative AI system were used to write an internally friendly email to the team, there would be no problems. However, if the public generative AI system is used to write a blog post on the official enterprise website, intellectual property, and copyright regulations should be considered, as these generative AI systems may copy text from other enterprise public websites.?

The final point is related to the overall process; using the generative AI systems is supposed to?“Boost the Productivity of the Enterprise”?and not to free the employee task list. Hence, the usage of these generative AI systems should be disclosed explicitly by the employees to their higher managers. Covert use of these systems to finish tasks while accounting for the same effort of the manual task is totally against the?“Professional Standards”?and?“Professional Integrity.”

Hence, and to conclude, there are many professional standards that should be followed when using public general-purpose lange model services, including the?“Input Data,”?“Output Data,”?as well as the?“Overall Process”?of using them. First,?the?“Well-Educated Employees”?should apply data protection and security measures. Second, enterprises need to manage the?“Challenging Paradox of AI Governance”?carefully.?Third, enterprises should be careful about the external use of the generative AI output data as it is usually subject to?“Intellectual Property and Copyrights Regulations.”?Finally, employees should explicitly disclose the usage of these generative AI systems to their higher managers as covert use of these systems to finish tasks while accounting for the same effort of the manual task is totally against the?“Professional Standards”?and?“Professional Integrity. These four guidelines ensure the minimum level of the?“Responsible Usage of Generative AI”?within the enterprise.?