When was the first ai robot made ？The first widely recognized AI robot was Shakey, developed at SRI starting in 1966 (the project ran from 1966 to 1972). Shakey could perceive its environment and reason about what to do next, which is the key difference between an AI robot and a purely automated machine.
One quick caveat: if by “first robot” you mean the first industrial robot used in a factory (not AI), that milestone is usually credited to Unimate, which went into service on a General Motors line in 1961.
This article explains why “first” gets messy, what counts as an AI robot, and where Shakey fits in the timeline.
Quick Answer: The First Widely Recognized AI Robot (with the Year)
Answer: Shakey — 1966 (research and development continued through 1972).
Why it counts as AI: Shakey wasn’t just following a fixed script. It combined sensing (like cameras and touch sensors) with planning and problem-solving so it could interpret instructions such as “push the block off the platform” and figure out a sequence of actions to make that happen.
If you’re writing a one-sentence “definition” for your memory: Shakey is often described as the first mobile robot able to perceive and reason about its surroundings.
What Counts as an “AI Robot”?
This is where most “first AI robot” debates start. People use “AI robot” to mean different things, so they land on different “firsts.”
Robot vs. AI Robot vs. Industrial Robot
-
Robot (broadly): A machine that can sense and act in the physical world.
-
Industrial robot: A robot built for repetitive, reliable tasks—often in manufacturing—typically operating in a controlled environment. Unimate is the classic early example.
-
AI robot: A robot that does more than repeat a preprogrammed routine. It uses perception + reasoning/planning + action in a loop—so it can cope with variation.
You can think of an AI robot as a system that answers four questions continuously:
-
What am I seeing? (perception)
-
What does it mean? (internal representation)
-
What should I do next? (planning/reasoning)
-
Did it work? (feedback and adjustment)
Shakey is historically famous because it pulled those pieces together—within the limits of 1960s computing.
Why “First” Depends on Your Definition
Even if everyone agreed on what “AI robot” means, the word made can mean:
-
Started (R&D began)
-
Demonstrated (first public demos)
-
Completed (end of a research program)
-
Deployed (used outside a lab)
-
Commercialized (sold broadly)
That’s why you’ll see different years attached to the same robot. With Shakey, it’s common to cite 1966 as the start of the project, and 1972 as the endpoint of the research program.
The Timeline: Key Milestones Leading to the First AI Robot
Let’s place Shakey in context, because it makes the “first” question easier to answer—and it helps you avoid mixing up “first industrial robot” with “first AI robot.”
Early Automation and the Rise of “Robots”
Long before anyone could build a robot that “thought,” engineers built machines that moved precisely, repeatedly, and safely. These systems were revolutionary—but they were usually automation, not AI: the machine did what it was told, and it didn’t do much interpretation.
That distinction matters because early factory robots changed manufacturing first, while AI robots grew out of research labs.
The First Industrial Robot (Not AI): Unimate (1961)
If your mental image of “the first robot” is a big mechanical arm doing dangerous factory work, you’re probably thinking of Unimate.
-
Unimate is widely credited as the first industrial robot used on a production line.
-
It went into service at a General Motors plant in 1961, handling hot die-cast parts—exactly the kind of job you don’t want humans doing.
Unimate is a milestone in robotics history—but it’s not usually called an “AI robot” because the intelligence piece (world modeling, goal-based planning, reasoning from perception) wasn’t the point.
The First Widely Recognized AI Robot: Shakey (1966–1972)
Shakey’s significance is that it was designed to be a general-purpose mobile robot that could interpret tasks, plan, and act. SRI describes Shakey as “the first mobile robot with the ability to perceive and reason about its surroundings,” and notes the research ran from 1966 to 1972.
Computer History Museum describes how Shakey used cameras and touch sensors, and how it combined computer vision, language processing, and planning to carry out commands like “push the block off the platform.”
The IEEE History Center (ETHW) frames it as “the world’s first mobile intelligent robot,” highlighting capabilities like planning, recovering from errors, and communicating using ordinary English.
Why Do Sources Disagree on the Date?
If you’ve seen articles claiming “1972” instead of “1966,” you’re not imagining it. They’re usually picking a different “timestamp” for the same story.
“Made” vs. “Developed” vs. “Demonstrated” vs. “Completed”
-
1966 is commonly used because that’s when SRI’s Shakey research program is typically dated as beginning.
-
1972 appears because it’s often cited as the end of the Shakey research period, and some historical writeups label the milestone around that year.
Both can be “right,” depending on what you mean by “made.” For a search query like “when was the first ai robot made,” the best user-centered answer is usually: Shakey began in 1966 (1966–1972).
AI Definitions Have Shifted Over Time
In the 1960s, a lot of AI work was symbolic AI: planning, logic, search, and structured representations of the world. Shakey sits squarely in that tradition.
Modern readers sometimes expect “AI robot” to mean neural networks or deep learning. But historically, “AI” in robotics originally meant something closer to: can the robot reason about actions in a changing environment? That’s exactly why Shakey still shows up in serious histories of AI and robotics.
What Made Shakey “AI”? A Capabilities Breakdown
It’s easy to say “Shakey was the first AI robot.” It’s more convincing (and more useful) to explain what it actually did.
Perception: Seeing and Sensing the World
Shakey was equipped with sensors such as cameras and touch/bump sensors, using them to detect parts of its environment.
Representation: Turning Sensory Data into “What’s Out There”
Raw pixels and sensor hits aren’t enough. The robot needed an internal model—an idea of rooms, objects, and relationships—so it could plan.
That’s one reason Shakey is remembered as a research platform: it forced AI researchers to bridge the gap between abstract reasoning and messy physical reality.
Planning / Reasoning: Figuring Out a Sequence of Actions
Shakey is closely associated with planning—the idea that you can give a goal and the system can generate steps to reach it.
The Computer History Museum description captures this well: Shakey could take instructions like “push the block off the platform” and direct its own actions using planning techniques.
IEEE’s milestone page also emphasizes planning, inference, and recovering from execution errors—concepts that are still central in robotics today.
Action: Moving Through Space and Manipulating Simple Objects
Shakey was mobile. It navigated around rooms and interacted with objects in relatively controlled settings—simple by today’s standards, but a huge leap at the time.
Conclusion
So, when was the first AI robot made? If you’re looking for the most widely accepted, history-friendly answer, it’s Shakey, developed at SRI beginning in 1966 (with the research program running through 1972).
That said, the word “first” can point to different milestones. If what you really mean is the first robot to make a real impact on factory floors, that honor typically goes to Unimate, which entered production-line use at General Motors in 1961—a landmark in industrial automation, even though it isn’t usually considered an AI robot.
The simplest way to keep it straight is this: Unimate was a “first” for industrial robotics; Shakey was a “first” for AI in robotics. If you’re researching early AI robots, start with Shakey’s core idea—perceive, plan, act—because that loop still defines what we mean by intelligent robots today.
FAQs
Was the first AI robot the same as the first robot?
No. “First robot” can mean many things—early automated machines, industrial robots, or research robots. The first widely recognized AI robot is often cited as Shakey (1966–1972), while the first industrial robot in factory use is typically credited to Unimate (1961).
Was the first industrial robot AI?
Not in the way “AI robot” is usually meant today. Unimate was a landmark industrial machine used on a GM line in 1961, but its historical importance is factory automation and safety—rather than perception-and-reasoning intelligence.
Why is Shakey considered AI?
Because it combined sensing with goal-based planning—it could interpret tasks and choose actions rather than following a rigid, prewritten sequence. That “perceive + reason + act” loop is a core idea behind AI robotics.
