Employees are already using AI tools such as ChatGPT, DALL-E 2 and Midjourney to do some of their work. 

They are using it at work (on your network), on their phones and at home. Although AI is not yet fully integrated into every piece of software used within the workplace, one should be under no illusion that employees are not using AI to help them do their jobs faster and more efficiently.

Many leaders are focused on incorporating AI into their products and services but they also need to think about how AI can be used now and in the future in the day-to-day operation of their companies. What is your AI plan for your company?

Create and communicate your AI use policy to employees

If you do not have an AI use policy in place, your first step is to meet with your legal team, information technology team and human resources team to get this process moving forward. Having basic guidelines for your employees to follow will be important. Your IT team may have strong opinions on which AI tools your company should or should not use. Your legal team may raise significant concerns over issues ranging from who owns the intellectual property generated from AI.to prohibited information that your legal team may not want inputted into an AI tool. Do not wait. Get ahead of this situation now.

Best practices for AI use in the workplace

The following best practices should be considered when using AI at work:

Use AI appropriately: AI should only be used for tasks that it is designed for and that are within its capabilities. It should not be used for tasks that require human judgment or decision-making.

Verify information: The information provided by AI should be verified before use to ensure its accuracy and reliability. This is one of the more critical aspects of employees using AI output for work. AI is not always right and can produce coherent but inaccurate information. It will in most cases be able to nail an uplifting email communication from the CEO to employees at Christmas but may run into issues if it is asked to write a technical white paper on an advanced microprocessor. Verification is critical.

Protect sensitive information: AI should not be used to generate or provide sensitive information, such as personal or financial data, employee emails, employee names or the company name, without appropriate security measures in place.

Maintain professional communication: AI should be used in a professional manner, and its responses should be respectful and appropriate for the workplace.

Monitor and evaluate AI use: AI use should be monitored and evaluated regularly to identify any unintended consequences or biases.

Ethical considerations

The use of AI must be guided by ethical considerations, including the following:

Fairness and nondiscrimination: AI should not be used to discriminate against individuals or groups on the basis of race, gender, age, religion or other personal characteristics.

Privacy and data protection: AI must be used in a way that respects individuals’ privacy rights and protects their personal data.

Transparency and accountability: AI should be transparent in its responses, and its outputs must be auditable and explainable. Those responsible for using AI must be held accountable for how they use it.

Assess every role in your company

The next step is to identify roles that can be automated using AI. It is important to analyze the tasks performed by each role and identify those that are repetitive, rules-based, content based and data-driven.

Assessing the impact of AI on roles will evolve over time, with some roles being impacted sooner than others:

1. Assess AI that is currently available for use today: Take into consideration those AI tools that are available within the public domain that your employees are using either with or without the company’s knowledge. For example, assess how ChatGPT could be used at present to increase productivity. (Zero to two years out)
2. Assess the tools that you use today to run your business: This is a good time to work with software and tools vendors and service providers to understand their AI roadmaps. Their roadmaps will give you a better understanding of how AI will be integrated into existing or future productivity tools used within your organization. These could range from financial systems, coding copilots, enterprise email, to HR systems. (Two to five years out)
3. Identify the tasks and activities associated with each role: Start by making a list of all the tasks and activities that are performed by each role in your organization.
4. Evaluate the suitability of each task for automation: Once you have identified the tasks and activities associated with each role, evaluate whether they are suitable for automation. Consider the complexity of the task, the amount of time it takes to complete, and the potential impact of automating it.
5. Assess the cost savings and benefits of automation: Determine the potential cost savings and benefits of automating each task. This could include reduced labor costs, increased efficiency, and improved accuracy.
6. Consider the impact on employees: Evaluate the impact that automation would have on the employees in each role. This could include job loss, reassignment of responsibilities, and upskilling opportunities.
7. Prioritize the roles for automation: Based on the results of your evaluation, prioritize the roles that are most suitable for automation. Consider the potential cost savings and benefits, as well as the impact on employees.
8. Plan for implementation: Develop a plan for implementing automation in the identified roles. This may involve selecting the appropriate AI technology, training employees on new tasks, and providing support for employees who are impacted by the changes.

It’s important to note that assessing roles for automation should not be solely focused on eliminating jobs. Rather, it should be about improving efficiency, productivity and accuracy, while also providing opportunities for upskilling and reskilling for employees.

Grade each role on AI’s impact on each role

One approach to understanding AI’s impact on each role is to try to forecast how existing AI.tools might impact each role and ultimately each department. For example, one could list every role within a department and grade how AI might have an impact on each role. A company’s grading system could vary from simple to complex, preliminary to extensive.

The chart below is a department-level rollup of various AI impact role assessments within each department. It assesses the impact from zero to two years and the impact of AI on department roles three to five years out.

Job elimination or reallocation?

The impact that AI will have on specific roles and organizations will vary depending upon the needs of the company. It will be up to your organization to fully assess if roles will be eliminated or reallocated.

Example 1: A software company using AI to develop more products at a faster rate may opt to reallocate engineers to work on designing new products and creating product requirements which will be “fed” into AI, as well as verifying that the AI output meets design requirements.

Example 2: A marketing agency using AI may opt to eliminate various roles within its company due to significant productivity increases with the use of AI.

Will college-educated and skilled workers be necessary?

A recent paper by the National Bureau of Economic Research titled “Generative AI At Work” by Erik Brynjolfsson, Lindsey Raymond and Danielle Li conducted research on the staggered introduction of a generative AI-based conversational assistant using data from 5,179 customer support agents. The study found that “AI assistance disproportionately increases the performance of less skilled and less experienced workers across all productivity measures we consider. In addition, we find that the AI tool helps newer agents move more quickly down the experience curve: treated agents with two months of tenure perform just as well as untreated agents with over six months of tenure.” In what other roles can AI level the playing field between less skilled and less experienced workers with more skilled and more experienced workers? 

The NBER’s “Generative AI At Work” study presents a fascinating view into some of the research being done on roles within the workplace. It is also an example of what each organization should do in terms of measuring productivity before and after the introduction of AI into various roles and departments.

CEOs will still need highly educated and skilled workers, but depending upon growth projections, they may not need as many. The role of your educated and skilled workers will lean toward being “AI Guides” and conducting “AI Validation.” In other words, some will be feeding AI requirements and others will be validating AI output. In some cases, they will be performing both tasks.

Conclusion

I disagree with the notion presented by the Center for Humane Technology and Elon Musk that humans can control AI, either practically or ethically. It’s improbable that AI will regulate or undo itself any more than the humans who are expediting its development will. The majority of individuals will use AI to their advantage, whether it is to enhance productivity or profit from it. AI will have a significant impact on every role within the workplace, making it our new colleague and our new reality. CEOs have no choice but to participate in the race, as AI shows no signs of slowing down, and those who don’t compete have already lost.

Joseph Hickman has worked in various prominent tech companies such as Oracle Corp., Google LLC and Tesla Inc. His previous articles for SiliconANGLE include “Your workers want to keep working from home. Now what?” and “Dear Twitter employees: Welcome to a new world.” He wrote this article for SiliconANGLE.

Images: Joseph Hickman/Midjourney