Albert Galarza is the Global Vice President of Human Resources at TELUS International.
getty
Generative AI (GenAI) is dominating headlines, and for good reason. For many, it may have felt like the technology came out of nowhere and rocketed into every domain, from art to education to business.
Its potential to cause paradigm shifts in the nature of work is of great interest, but there’s also some cause for concern. Employers are apprehensive about whether employees should leverage GenAI in their work, with some big companies going so far as to ban its use outright for fears of data leaks, plagiarism and infringements on trademarks and intellectual property. The fact that GenAI remains prone to misinformation also raises questions about its accuracy and effectiveness.
As an emerging technology, GenAI can offer significant benefits for brands that integrate it into their business operations. But effective adoption requires organizations to take a responsible approach and the necessary steps to safeguard against harm and build trust within their workforce.
5 Tips For A Thoughtful, Responsible Approach To Generative AI
As your organization begins integrating GenAI into its systems, it’s crucial to get it right from the start. There are a number of critical factors to consider so you can ensure accuracy and fairness. Here are five areas you should focus on.
1. Use diverse and representative data sets.
The foundation of any GenAI system lies in the training data. Algorithms create new outputs based on the data they've been trained on, magnifying the impact of data quality on overall performance. So to help mitigate bias in the outputs, you should ensure the data sets used are diverse and representative. By incorporating data from a wide range of sources and demographics, you can reduce the risk of a GenAI system perpetuating biases or creating skewed outputs. However, it’s not a one-time deal. Regularly evaluating and updating data sets is a must to maintain accuracy and fairness over time.
The need for good and fair data has become more important than ever. In fact, research by my firm, TELUS International, uncovered that 77% of American consumers believe GenAI algorithm audits around bias and prejudice should be mandatory before brands integrate this technology into their platforms.
2. Establish policies and guardrails.
Guardrails act as constraints that help set acceptable boundaries to keep GenAI systems in check. From topical and safety guardrails to security guardrails, crucial policy helps control and influence the outputs of GenAI. For example, you may want to clearly define the appropriate use cases for this technology, like when employees are conducting research or communicating externally with customers, to ensure compliance with privacy laws, client agreements and corporate standards. By implementing appropriate policy guardrails, you can better define the controls that may protect your organization against the most critical risks of GenAI tools, including data leaks and toxic or harmful content.
3. Engage in continuous monitoring and evaluation.
Establish mechanisms for continuous monitoring and evaluation of your GenAI systems. Model validation and fine-tuning are crucial because a key function of GenAI tools is enhancing employees' decision-making. By leveraging advanced algorithms, your organization can collect valuable insights and analyses that aid in making strategic, data-driven choices. For instance, HR professionals can use AI-powered systems to optimize talent acquisition processes and identify top candidates efficiently.
Regular audits and reviews should assess the system's performance, identify potential issues and address biases or misinformation. Then, you can make any necessary adjustments to data or processes to improve accuracy and ensure the system effectively adapts and makes accurate predictions. Employing techniques such as data anonymization and data augmentation can help identify potential biases or misinformation and reduce their impact on the GenAI system's outputs.
4. Incorporate a human-in-the-loop approach.
Human feedback is necessary for GenAI systems to operate responsibly and ethically. People play a crucial role in reviewing and validating the outputs of AI-based systems, as they can flag any biased or inaccurate content, check for appropriate context, identify potential pitfalls and make informed decisions in cases where the system's outputs require additional scrutiny. So ensure you're incorporating insights from the people who are working directly with the systems on a daily basis.
GenAI-powered tools can capture and categorize organizational information, facilitate cross-functional collaboration and enable knowledge sharing among teams. This makes data produced by more experienced employees and top performers incredibly valuable as part of the data set that continuously trains systems. This can also be particularly helpful in reducing nesting time for newer employees or those requiring additional support.
5. Ensure transparent communications and disclosure.
Finally, you should prioritize transparency and disclosure regarding your organization's use of GenAI systems. Transparency has always been a foundational block of successful relationships—both professional and personal—so clear communication with stakeholders, employees and end-users about the presence and purpose of AI-generated content is essential. Sharing information about the model's limitations and potential biases, as well as your ongoing efforts to mitigate them, also helps build trust and accountability.
Looking Ahead
Generative AI technology has the potential to vastly change and enable the way we work. So the way you approach integration is vital. By taking the necessary steps to ensure GenAI-enabled platforms are accurate, safe and sustainable, you'll be well positioned to reap the rewards.
Forbes Human Resources Council is an invitation-only organization for HR executives across all industries. Do I qualify?
