In a new development that some may find disconcerting, researchers in the US have made a knife-wielding robot smarter and capable of making its own decisions.

The robot, being developed for food-handling duties such as supermarket check-out work or as an in-home chef, needed the extra smarts to make sure it knew what it was handling and how to stay safe.

The update brings the possibility of robot butlers even closer, but also the likelihood of future cyborg replacements for supermarket staff. Luckily, they will still need some on-the-job training before pushing people from their positions.

Engineers at Cornell University have taught their android to work in a mock supermarket check-out line, modifying the robot to “coactively learn” from humans, making adjustments while an action is in progress.

“We give the robot a lot of flexibility in learning,” said Ashutosh Saxena, assistant professor of computer science.

“The robot can learn from corrective human feedback in order to plan its actions that are suitable to the environment and the objects present.”

Creating robots for assembly lines has been relatively simple. Once a particular action has been programmed, the unit only needs to repeat it until it is told to do something else.

In real life, at the supermarket or in the home, things are much more complicated – so robot capabilities need a lot more programming.

An android in the kitchen would need to know how to handle tomatoes, eggs, soft and delicate foods, as well as wield knives and other implements in a safe manner around humans.

The updates are designed to allow more cooperation and just a little bit of initiative for the future domestic helpers.

The robot can plan some of its own movement trajectories, which are then presented to the operator, who can pick the one that seems the most effective and/or safest.

The learning algorithm the researchers provided allows the robot to learn incrementally, refining its trajectory a little bit more each time the human operator makes adjustments or selects a trajectory on a touch screen.

Even with weak but incrementally correct feedback from the user, the robot arrives at an optimal movement.

Robots were programmed to learn to associate a particular trajectory with each type of object.

A quick flip over might be the fastest way to move a cereal box, but that could be disastrous with a carton of eggs. Also, since eggs are fragile, the robot is taught that they shouldn't be lifted far above the counter. Likewise, the robot learns that sharp objects shouldn't be moved in a wide swing; they are held in close, away from frail humans.

In tests with users who were not part of the research team, most users were able to train the robot successfully on a particular task with just five corrective feedbacks. The robots also were able to generalise what they learned, adjusting when the object, the environment or both were changed.

A video demonstration of the updates is available.