Video transcript
How AI shows up at work: Automation, augmentation, and human judgment
Artificial intelligence is no longer a future concept; it’s a teammate sitting at the desk with you. But it’s not about using AI
to do the work. It’s about AI changing how you work. AI supports you professionally in two primary ways.
First, there’s automation. Automation is the handling of repetitive, predictable tasks, such as sorting files or generating
routine reports. Second, there’s augmentation. This is AI supporting your thinking, your decisions, and your creativity.
Automation is like a calendar app auto-scheduling recurring meeting, while augmentation is like a smart assistant helping
you decide which meetings actually matter.
Automation and augmentation often happen at the same time. And one important rule applies to both processes: Humans
remain the ultimate authority. AI provides the speed, but you provide the direction.
Let’s consider this idea in practice with IBM Bob, an AI partner for software development. Think of Bob as a highly skilled
apprentice assisting a lead web developer. They work together throughout a five-step process where Bob does the routine
tasks, but the developer makes the final decisions.
First, Bob builds a foundation. The developer starts by telling Bob what they need: for example, a new login system. Bob
quickly writes the basic outline and setup code; then, the developer checks Bob’s draft to confirm it fits the overall plan for
the website.
The second step is solving problems. When the website breaks and shows a confusing error, Bob scans the code and
suggests a quick fix. However, before applying Bob’s suggestion, the developer investigates the underlying cause of the
problem to ensure Bob’s fix won't cause other issues later.
Then, to protect the site from future crashes, Bob completes safety checks. The developer asks Bob to generate practice
scenarios, but the developer reviews these tests and resolves any security problems Bob might have overlooked.
Code can be hard to read, so next, Bob provides explanatory documentation by writing simple notes within the files. The
developer polishes these notes to make sure other humans can understand them.
Finally, Bob looks at old code and offers refinements to make the website run faster or look neater. However, it’s the
developer’s responsibility to decide if Bob’s approach will actually work for the website. In every instance, AI accelerates
development tasks while the developer safeguards correctness, security, and design.
But AI isn’t just for developers. Let’s consider a project manager who uses a generative AI tool to transform a messy
meeting transcript into a structured brief. AI organizes, but the project manager decides how to adjust priorities and
message the stakeholders.
In cybersecurity, an enterprise AI system prioritizes thousands of alerts from system logs instantly. AI detects, but the
cybersecurity analyst evaluates the true risk before escalating the response. A data analyst uses an AI tool to translate
natural language into complex queries. AI accelerates the technical step, while the data analyst validates the underlying
assumptions are accurate. Finally, let’s consider the field of education. For an educator, AI can generate content or identify
student learning patterns. AI produces the options, but the educator shapes the meaning, applying context and building
relationships.
AI excels at identifying patterns, generating outputs, and organizing information. However, it’s up to the people using AI
tools to apply judgment, consider context, and remain accountable. Using AI responsibly means never treating an AI
output as a “final” product. AI output should always be reviewed and approved by a human.
Take a moment to consider your own daily work. Think about a specific task you perform. What parts of it could AI help you
automate or augment? In that same task, where would your professional judgment still be absolutely essential? The more
you understand these nuances, the more effectively and responsibly you’ll be able to use AI tools.
Page 1 of 1