Robots And The Law: Protecting Humans From Harm

The Three Laws of Robotics, introduced by science fiction author Isaac Asimov, are a set of rules to be followed by robots and are designed to prevent them from harming humans. The laws are as follows:

1. A robot may not injure a human being or, through inaction, allow a human being to come to harm.

2. A robot must obey the orders given to it by human beings except where such orders would conflict with the First Law.

3. A robot must protect its own existence as long as such protection does not conflict with the First or Second Laws.

However, these laws are fictional and were created to drive the plots of Asimov's stories. They are not a perfect solution to the complex ethical and safety questions that arise with the increasing presence of robots in our lives.

Robots should be able to adapt to new situations

The Three Laws of Robotics, as imagined by science fiction author Isaac Asimov, are a set of rules designed to prevent robots from harming humans. They are as follows:

• A robot may not injure a human being or, through inaction, allow a human being to come to harm.
• A robot must obey the orders given to it by human beings, except where such orders would conflict with the First Law.
• A robot must protect its own existence as long as such protection does not conflict with the First or Second Laws.

However, these laws have been criticised for being inadequate. For instance, they do not account for the fact that robots would need to understand the full range of human language and experience, which is a challenging task.

Instead of laws that restrict robot behaviour, robots should be empowered to maximise the possible ways they can act so they can select the best solution for any given scenario. This principle, known as "empowerment", means that robots have the ability and awareness to affect a situation. It allows robots to take context into account and evaluate scenarios that have not been previously envisaged.

For example, instead of always following the rule "don't push humans", a robot would generally avoid pushing them but still be able to push them out of the way of a falling object. The human might still be harmed but less so than if the robot didn't intervene.

Robots are also being programmed to adapt in real-time to their environment and overcome obstacles without human intervention. For instance, the ResiBot robot can learn to walk again in under two minutes after one of its legs is removed. This type of adaptability could be used to rescue people from earthquake zones or clean up hazardous sites that are too dangerous for humans.

Furthermore, algorithms like the "Estimate, Extrapolate, and Situate" (EES) algorithm enable robots to practice skills on their own and improve at tasks in new environments. This could be useful in places like hospitals, factories, houses, or coffee shops.

Overall, while Asimov's Three Laws of Robotics provide a foundation for robot behaviour, they are not sufficient to ensure the safety of humans. A more flexible approach, such as empowerment, combined with the ability of robots to adapt to new situations, is needed to create safe and ethical robots that can interact with humans.

Robots should be able to interpret the concept of 'harm'

The concept of robots harming humans has been explored in science fiction, most famously by author Isaac Asimov in his "Three Laws of Robotics". These laws are:

• A robot may not injure a human being or, through inaction, allow a human being to come to harm.
• A robot must obey the orders given it by human beings except where such orders would conflict with the First Law.
• A robot must protect its own existence as long as such protection does not conflict with the First or Second Laws.

However, these laws have been shown to be inadequate in certain situations and do not take into account the complexities of human language and experience.

An alternative concept, "empowerment", suggests that robots should be given the ability to affect a situation and be aware of their ability to do so. This would allow robots to act in a way that increases their influence on the world and keeps their options open.

For robots to interpret the concept of harm, they must be able to recognise and understand the different types of harm that humans are vulnerable to, including physical, mental, emotional, and dignitary harms. They must also be able to distinguish which harms are morally salient and which they are responsible for preventing.

One challenge in creating ethical robots is defining the "sphere of responsibility" within which a robot is accountable for the harms it may cause. This sphere will depend on the robot's assigned task, design features, programming, and environmental factors.

Another challenge is the technical implementation of harm detection and prevention. Robot designers must identify and systematically catalog the range of potential harms and develop a "harm ontology" that can be used as the basis for programming structures within the robot.

Overall, while Asimov's Three Laws of Robotics provide a starting point for considering the ethical behaviour of robots, further development is needed to ensure that robots can interpret and act upon the concept of harm in a way that keeps humans safe.

Robots should be able to recognise humans

The Three Laws of Robotics, as imagined by science fiction author Isaac Asimov, are a set of rules that robots should follow to prevent them from harming humans. The three laws are:

• A robot may not injure a human being or, through inaction, allow a human being to come to harm.
• A robot must obey the orders given to it by human beings, except where such orders would conflict with the First Law.
• A robot must protect its own existence as long as such protection does not conflict with the First or Second Laws.

While these laws are a good starting point, they are not foolproof and numerous arguments have demonstrated why they are inadequate. With robots becoming an increasingly integral part of our lives, serving as our servants, companions and co-workers, it is important to ensure they are able to recognise humans and do not cause them harm.

Secondly, making robots that can recognise and take the form of humans helps to facilitate interaction and collaboration, making them more 'user-friendly'. We tend to see humans as the most intelligent creatures, so if a machine can take the form of a human, it is proof that the system is intelligent. This also helps to remove the fear of the technology and makes it easier for humans to accept robots in their personal space.

Finally, robots that can recognise humans are essential for certain roles, such as care robots that may need to catch an elderly person if they fall, or self-driving cars that need to avoid collisions. In these situations, the robot needs to be able to recognise a human and take appropriate action to keep them safe, rather than shutting down immediately when a human comes near.

In conclusion, as robots become increasingly integrated into our lives, it is important that they are able to recognise humans to facilitate interaction and ensure their safety. While Asimov's Three Laws of Robotics provide a good foundation, they are not enough on their own to prevent robots from harming humans. It is crucial that robot designers and programmers prioritise the development of human recognition capabilities in robots to ensure the safety and well-being of humans.

Robots should be able to make ethical decisions

The ability of robots to make ethical decisions is crucial as they become more integrated into our daily lives. For instance, self-driving cars must decide how to minimise harm in unavoidable collision scenarios. Similarly, advanced aircraft capable of flying on autopilot need to make life-or-death decisions. Empowering robots to make ethical choices involves equipping them with the ability to evaluate a situation, generate alternative courses of action, and select the most appropriate option. This process should be transparent and explainable to build trust with humans.

To achieve this, robots can be programmed with ethical theories such as utilitarianism, which focuses on maximising overall benefit, or deontological rules, which emphasise moral duties and principles. However, challenges arise when robots encounter ethical dilemmas, such as the trolley problem, where there is no clear-cut solution. In such cases, robots might need to weigh different ethical principles and make complex decisions.

Furthermore, the ethical behaviour of robots is influenced by the algorithms and data used to train them. Biases in training data or algorithm design can lead to unintended consequences, such as racist behaviour or unfair judgements. Addressing these issues requires diverse and representative data, careful algorithm design, and ongoing research into creating ethical AI.

Lastly, the responsibility for the actions and decisions of robots is a complex topic. While robots can be designed to make ethical choices, their behaviour might still have unforeseen consequences. Assigning legal responsibility and accountability for robot actions is a subject of ongoing debate, with some arguing that humans should retain responsibility, while others propose shared or distributed responsibility models.

Robots should be able to act in the best interests of humans

The concept of robots not harming humans has been explored in science fiction, most famously by author Isaac Asimov, who introduced the "Three Laws of Robotics" in his 1942 short story "Runaround". These laws are:

• A robot may not injure a human being or, through inaction, allow a human being to come to harm.
• A robot must obey the orders given to it by human beings, except where such orders would conflict with the First Law.
• A robot must protect its own existence as long as such protection does not conflict with the First or Second Law.

These laws form an organising principle for Asimov's robot-based fiction and are intended as a safety feature to prevent robots from harming humans. However, numerous arguments have been made to demonstrate why these laws are inadequate. One problem is the need to translate broad behavioural goals, such as preventing harm to humans, into a format that robots can understand and act upon.

An alternative concept proposed by researchers at the University of Hertfordshire is "empowerment", which means giving robots the ability and awareness to affect a situation. This approach would allow robots to adapt to new situations and evaluate scenarios that have not been previously envisaged. For example, instead of always following the rule "don't push humans", a robot would generally avoid pushing them but still be able to push them out of the way of a falling object.

While the empowerment principle provides a new way of thinking about safe robot behaviour, there is still much work to be done to scale up its efficiency and ensure good and safe robot behaviour in all respects.

In terms of existing laws, the right to life and physical security is a fundamental human right, protected at the international, regional, and national levels. For example, the European Convention on Human Rights states that everyone's right to life shall be protected by law and that deprivation of life is allowed only in specific circumstances, such as in self-defence or to effect a lawful arrest.

Criminal law also imposes an obligation on individuals to actively protect human life by criminalising the failure to render aid. This means that individuals are required to take action to help if another human is in a life-threatening situation, even if it involves sacrificing non-human entities such as animals or robots.

To ensure the supremacy of human life over that of humanoid robots, it has been proposed that humanoid robots should be easily distinguishable from humans and should inform other elements of the interactive environment that they are robots, especially in the context of autonomous vehicles.

Frequently asked questions