AI and Liability, who will pay for machine mistakes?

You will learn: who will be responsible for damages caused by AI errors; Is there a division of responsibility between developers, users and other stakeholders; Let’s consider the mechanisms for protecting oneself from liability and what legal norms apply in the above cases.

Life is getting easier, we can delegate more and more tasks to AI. The future seems to have arrived right now. But is everything so easy and “sweet”?

AI is evolving and becoming more accessible to us. Now you can use it to do almost anything: from creating a website without programming knowledge to interfering with complex systems with the help of a created virus. What seemed unreal a few years ago is now in front of us in a browser tab.

AI can be described as a complex system that includes software code and a database. For their interaction, there are algorithms that analyze information and perform data processing, performing functions similar to human mental processes, such as language recognition, pattern recognition and decision-making.

But AI cannot work in a vacuum, it exists based on a database. It is these databases that serve as a source of information for training AI. Thanks to this, the system adapts and improves over time.

• If you think that you have not used AI before, because you have never used such services, I will surprise you: artificial intelligence is everywhere.

In finance, social media, advertising, research and technology, and even on your smartphone.

When you unlock your phone with Face ID or use speech recognition, you’re talking to nothing but AI.

AI exists, its prevalence is clear, but who is responsible?

And now a surprise: there is no proper legal regulation, it is just beginning. If now you have a question: what will we talk about in this article? Don’t worry, there’s plenty to discuss here.

AI is not legally regulated, so in this case, we act according to the principle that everything that is not prohibited is allowed!

This kind of permissiveness, at the same time it is important to understand that the law does not have retroactive effect, which means that any newly adopted law does not apply to the actions of the past. That is, it is not possible to punish actions that were legally permitted in the past, but now fall under regulation.

But still, AI does not exist in a legislative vacuum, and therefore, according to the general understanding, responsibility for the actions of AI will be borne by those who create, implement and use these systems.

It is legally defined that the responsibility lies with the developers according to the principle of product liability. If the damage is found to be caused by defects in the product, the developer is responsible.

It is important to understand that the use of artificial intelligence also entails rights, duties and responsibilities. But due to the imperfection of the law, many questions remain unanswered in matters related to the use of AI and are resolved by analogy with other similar legal relations.

In order for you to understand what we are talking about, here is an example:

Tesla car owners often try to blame the company for accidents when they use the Autopilot system. That is, they are trying to shift the responsibility for damage caused by the actions of AI to the actual developers of this AI (in this case, the autopilot).

Here there is a situation when there is no legal regulation, and the situation itself already exists. There are several possible cause-and-effect relationships that can provoke this situation: shortcomings of the program itself; incorrect use of the autopilot by the driver; the intervention of third parties who, for example, hacked and damaged the program or made certain changes to it and, accordingly, the fault of such persons.

Does it work? Due to established case law in the US and good lawyers, Tesla often wins in courts during such cases.

Why? Their main arguments are quite logical – drivers should always pay attention when using autopilot and be ready to take control. (in fact, these warnings refer to the basic rules of the road, which in no way cease to operate with the use of autopilot, but on the contrary only reinforces it)

By the way, cars with activated autopilot have a lower frequency of accidents compared to those driven by drivers without such a system.

It is important to understand that users of these systems have their share of responsibility when using them, especially if they ignore the manufacturer’s recommendations or use the product in a way that is not intended. It is in such cases that the company “washes its hands” – it was used incorrectly, what can be the claims? After all, everything is written in the rules.

So the user can also be held responsible? So!

Let’s imagine that someone uses artificial intelligence to create and spread fake news or fraudulent messages. At the same time, the technology is intended to generate texts for educational purposes, using it for cheating is a clear abuse.

In this case, the responsibility falls on the user who abuses the technology, as well as on those who ignored laws and ethical norms when using artificial intelligence.

AI companies may even be required to release user data to law enforcement based on legitimate requests, including correspondence and other information related to users. Because, as I said earlier, AI does not exist in a legislative vacuum, and therefore all other laws do not cease to apply during its use.

Summarizing the information we have analyzed

Manufacturers and developers, in most cases, are considered the primary responsible parties. They have the primary duty to ensure the reliability and safety of the artificial intelligence system.

At the same time, users must not use AI contrary to existing laws and may also be held liable for fraudulent actions related to the use of AI.

These legal recommendations were developed by a lawyer in the field of protection of business interests and intellectual property. For professional legal advice, contact via Telegram: @your_legist.