Nine Ways that Managing an AI is Like Managing a Human, and Two Ways it’s Different
This article originally ran in insideBigData on May 29, 2018 (link here).
We’re not yet at the point where an Artificial Intelligence (AI) can storm into your office demanding a pay rise. But, managers are already expressing concerns that they are not ready for the AI revolution and don’t know how to manage AIs.
This blog post is a quick start guide for managers that shows how managing AIs compares and contrasts to the more familiar task of managing human staff.
Managing an AI Can Be Similar to Managing a Human
1. Job Descriptions
Before you even hire a new staff member, you need a job description for applicants. That job description describes the nature of the role, the responsibilities and authorities, and where the role fits within the organization.
Similarly, before you create an AI, you create a project statement. That project statement will outline the business problem to be solved, the decisions that the AI will make, and where the AI fits within the organization’s work flows and system architecture.
2. Interviewing and Recruiting
When recruiting new staff, a short list of suitable applicants will be selected for interviews. During those interviews, applicants will be asked to demonstrate their competencies and suitability for their desired role. The best applicant is offered the role.
Similarly, when choosing an AI, you will select a short list of algorithms that seem the most suitable. Those algorithms are given the chance to learn from your data, and then they need to prove their worth via head-to-head comparisons on accuracy, speed, and model insights. The best AI algorithm will go into production.
3. Key Performance Indicators (KPIs)
Each role in an organization will have key performance indicators. For example, a salesperson might have KPIs for number of new sales leads, and the total value of sales closed. KPIs are used to focus the staff member on how they add value to the organisation.
Similarly, an AI will have model metrics against which its performance is measured. These model metrics may include accuracy, speed, and insights.
4. Performance Reviews
An organization will hold regular performance reviews for each staff member. Successful staff need confirmation that they’re doing well, and less than successful performers need to understand the gap between what they’re delivering, and their organization’s expectations. Organizations need to know who gets promoted, who gets a big raise, who stays, and who goes.
Similarly, an AI’s performance must be monitored. Even the best AI will deteriorate over time, as the world around it changes. And then the organization needs to evaluate and decide whether to replace it or retrain it.
AIs and human employees both require your organization to spend time thinking about the best fit for your long-term goals.
5. Staff Training and Continuing Professional Development
Business environments are constantly changing. New technologies, legislative changes, and changes in consumer expectations require an organization to change work flows and restructure roles. Staff training prepares staff members for changes in work flows and roles.
Similarly as time goes on, the data on which an AI was trained becomes more and more obsolete. It is prudent to regularly schedule and retrain an AI on the latest data. This is called a “model refresh”.
Staff remuneration is structured to attract staff, reward good performance, retain good staff, and to keep costs affordable.
While you do not need to pay your AI in money, it does need system resources. If you are working your AI harder, then ensure that it has the system resources to meet your demands.
7. Human Resources Policies
Human resources policies exist to document company policies and procedures. They cover a wide variety of areas such as: standards of conduct, compensation, performance appraisal systems, employee benefits, and employee safety.
While many of these policies and procedures do not directly apply to AIs, applicable business rules need to be documented and applied, such as: data access, data validity rules, and decision authorities.
8. Regulatory Issues
There are many regulatory issues that apply to employing humans, including: minimum wages, occupational safety, discrimination, termination, annual leave, medical leave, and work visas.
The laws that apply to human employment do not apply to AIs.
However, regulations do apply to what an AI does. For example, your organization should not make discriminatory decisions, including decisions made by AIs. And data protection and privacy protection regulations are beginning to explicitly cover decisions made by AIs. For example Europe’s GDPR regulations provide consumers with the right to a review of an algorithmic decision.
When you hire staff, you have high hopes that you have made a good decision and that the employment relationship will be a long and mutually beneficial one. Sometimes however, it doesn’t work out as well as expected, and you may choose to terminate that person’s employment.
On the other hand, it is expected that you will replace AIs more frequently, as technology improves, or as better algorithms become available. This does not come with the complexity of terminating human staff and is simply a matter of version control.
How Managing an AI is Different to Managing a Human
1. Career Planning
Humans have a working life of approximately half a century. They start in junior roles, and as they develop skills and knowledge, they move into more senior roles. Managers will mentor staff in their careers, and human resources teams will carry out succession planning to ensure that there are staff who are ready to step up as new opportunities arise.
In contrast, AIs are specialists doing just one thing. They don’t develop into generalists, nor are they given management roles. And, with the rapidly changing developments in technology, AIs have a much shorter working life, unlikely to exceed several years at most.
2. Employee Relations and Retention
Humans are not just cogs in a machine. They are emotional, and require appreciation, recognition, communication, resolution of grievances and conflicts, and clear paths for career growth.
In contrast, AIs are machines. They will do their jobs the same every day as long as you provide them with the system resources.
AIs and human employees both require your organization to spend time thinking about the best fit for your long-term goals. And just as with humans, if you manage your AIs well, you can expect a long, and mutually-beneficial, relationship.
About the Author:
Colin Priest is the Director of Product Marketing for DataRobot, where he advises businesses on how to build business cases and successfully manage data science projects. Colin has held a number of CEO and general management roles, where he has championed data science initiatives in financial services, healthcare, security, oil and gas, government and marketing. Colin is a firm believer in data-based decision making and applying automation to improve customer experience. He is passionate about the science of healthcare and does pro-bono work to support cancer research.
VP, AI Strategy, DataRobot
Colin Priest is the VP of AI Strategy for DataRobot, where he advises businesses on how to build business cases and successfully manage data science projects. Colin has held a number of CEO and general management roles, where he has championed data science initiatives in financial services, healthcare, security, oil and gas, government and marketing. Colin is a firm believer in data-based decision making and applying automation to improve customer experience. He is passionate about the science of healthcare and does pro-bono work to support cancer research.
We will contact you shortly
We’re almost there! These are the next steps:
- Look out for an email from DataRobot with a subject line: Your Subscription Confirmation.
- Click the confirmation link to approve your consent.
- Done! You have now opted to receive communications about DataRobot’s products and services.
Didn’t receive the email? Please make sure to check your spam or junk folders.
How MLOps Enables Machine Learning Production at ScaleMarch 23, 2023· 4 min read
How the DataRobot AI Platform Is Delivering Value-Driven AIMarch 16, 2023· 4 min read
Discover insights on the specific conditions that make machine learning effective in certain financial applications, such as high-frequency trading. Read more.
In this article, we’ll first take a closer look at the concept of Real Estate Data Intelligence and the potential of AI to become a game changer in this niche.
In this blog post we’ll explore how Mindshare, a global media agency network, has leveraged data science tools to create a fast and reliable decision-making engine. Read more.