5 levels of robot autonomy: From production tapes to replacing people – how do machines take on more and more tasks?

chiny24.com 2 weeks ago

Robots have revolutionized manufacture for decades, but only now, thanks to artificial intelligence, are they starting to show real autonomy. According to the analysis SemiAnalysis of July 2025, developed by experts including Reykjav Knuhtsen and Dylan Patel, machines enter an era in which they will be able to replace people in almost all physical work. The key to knowing this revolution is the concept “Levels of Robot Autonomy” – a five-stage scale showing how far we have come and where we are going.

Level 0: Programmed movements – the era of "dark factories"

Level robots 0 These are classical industrial machines that only operate under strictly controlled conditions. They are highly precise, but they deficiency any independence.

Examples:
- Welding robots in car factories (e.g. Tesla, Toyota).
- Automatics in electronics factories (e.g. Foxconn mounting iPhones).
Data:
- In the car mill even works 1000 robotsand their cost is returned to 2 years.
- “Dark factories” (e.g. FANUC plant in Japan) operate without human participation – machines produce another robots making 80 seconds.
Challenges:
- Lack of flexibility – all change requires a full production line (cost even USD 60 million).

“It’s like building a cathedral – the task takes years and all modification is expensive” – comments anonymous automotive engineer.

Level 1: Smart choice and place – robots that see

With device learning algorithms, robots learned recognize objects and catch them with different strength and precision.

Examples:
- Sorting packages in Amazon magazines (e.g. Covariant robots).
- Logistics automation in courier companies (DHL, FedEx).
Data:
- Robots accomplish 99% efficacy in catching items, but jump to 99.99% required 81 times more data.
- In warehouses, 10 robots can replace 23 employees, reducing costs by 73%.
Challenges:
- Some objects (e.g. transparent bottles) inactive origin problems for robots.

Level 2: Autonomous mobility – robots in the open world

Level robots 2 can Navigate yourself in a dynamic environment specified as construction sites or refineries.

Examples:
- Spot from Boston Dynamics – inspection of oil platforms.
- ANYmal robots at atomic power plants.
Data:
- Autonomous inspections may save $500,000 per hour. (avoidance of downtime in refineries).
- Robots can work 90 minutes per batteryand their implementation continues 1 to 3 weeks.
Challenges:
- Navigation errors (e.g. confusing puddles with asphalt).

Level 3: Narrowly specialized manipulation – kitchen and laundry robots

Robots begin to execute simple work requiring skill, e.g. cooking or folding clothes.

Examples:
- Company CloudChef He's investigating robot cooks in restaurants.
- Watney Robotics automates industrial laundry.
Data:
- In San Francisco the rotation of gastronomy workers reaches 170% per year – robots are cheaper than continuous training.
- Robots can work in a model ‘Robot-as-a-Service’ (hourly fee as per employee).
Challenges:
- Slow pace of operation (e.g. presently folding towel takes A fewer minutes.).

Level 4: Tasks requiring “feeling” strength – last barrier

This is the future of robotics – machines that understand the force of pressure and execute precise tasks, e.g. installation of electronics or repair of plumbing.

Examples (research):
- Touch sensors (e.g. project) "Feel The Force").
- Manipulation with delicate objects (e.g. screwing screws into narrow slots).
Challenges:
- Lack of adequate data to train AI in the field of “feeling”.

“It’s like teaching a robot not to crush a strawberry, but at the same time it grabs a screwdriver” – explains Dr. Niko Ciminelli, co-author of the report.

When will robots replace humans?

According to forecasts:

2025-2030: Dissemination 3 (e.g. cleaning robots, cooks).
After 2030: Level input 4What could mean mass automation even 70% of physical professions.

The robotic revolution will not be abrupt – it is gradual taking over the next tasks. present in warehouses and factories machines work alongside people, but in a fewer years they can become the main labour force. The key question is: Are we ready for this?

Source: