Can AI really drive humanity to extinction?

We first heard that an artificial intelligence system “gained consciousness and feels,” according to a Google engineer. Later images of Dalle-e went viral and we entertained ourselves with ChatGPT. Then came the warnings, the fears, the requests for regulation.

And the doubts.

So at BBC Mundo we asked you a few days ago what you wanted to know about artificial intelligence (AI), we compiled your doubts and consulted them with an expert who has been working in the field of AI for more than 30 years.

This is Amparo Alonso Betanzos, professor of Computer Science and Artificial Intelligence at the University of A Coruña (Spain) and assistant to the rector for AI issues. She was also president of the Spanish Association of Artificial Intelligence (AEPIA).

This is what he told us.

How does it work and what is artificial intelligence like?

It’s hard to say because it has so many sub-areas, but there are basically two ways to approach artificial intelligence. One is the symbolic, the old AI, where we acquire knowledge from experts in the field, it is much more transparent but it is not scalable.

The other, the AI ​​that we have today, is based on data. In order to derive knowledge, what is done is to feed the system with data from a certain field, the system learns from that data and extracts the patterns. It is capable of generalizing, predicting, etc. in many areas, from natural language, computer vision or machine learning.

There are models where the process is done by reasoning based on deep learning with neural networks with many layers that end up learning that data. But there are other models such as reinforcement reasoning or other types that can be used to learn and derive knowledge to AI.

The AI ​​”feeds” on data. Where do those databases come from?

It depends a lot on the system you have. If it is an expert system for medical patterns, they are extracted from the large clinical databases that are referenced by certain types of diseases or certain types of patients. If it is traffic data, the available cameras or traffic sensors will be used.

Today the digitization process we are in is so immense that there are sensors that can derive data from practically any natural or industrial process that we can think of. Virtually all the experiences one can think of are digital: your trips, your medical records, your preferences…

For example, when you sit in front of the television and it recommends what to watch, it is based on what you have done before on that platform. Sometimes all this is the food of AI algorithms.

How is the AI? There are many who imagine a great computer or that it can become a T100-style machine from Terminator.

No. Not so… Unless you have your AI programs embedded in some anthropomorphic-looking robot. It can be like an automated vacuum cleaner that goes around the house or have a humanoid shape, but it is also simply that you turn on your computer and have software that listens to you, or a program on your mobile phone that detects your fingerprint.

It is impossible to say a number, how many there are. There are many systems and they are used for very different things, from the television that recommends what to watch to an app to predict if your vineyards will have a disease. It is something very transversal. It’s pretty much in any area you can think of.

And what is the impact of AI in our daily lives, in jobs…?

Many times we are using artificial intelligence and we are not even aware of it.

In the future we will tend to have more AI because it is being implemented in more and more areas.

Regarding employment, before the pandemic we saw how the panorama changed. There are many more jobs affected by automation, not just by AI. We see it in supermarkets, with more and more machines instead of cashiers, for example.

It will change the ways of working, especially in automated tasks, and we will have to live with the fact that part of our routine tasks is done by machines. I give you the example of doctors, who worked with hardly any instruments 50 years ago and today have many more machines at their disposal.

It is clear that it will affect jobs and the economy, and it is something that governments must deal with. We must be aware because, if not, it can create large unevenness. And yes, some jobs will be destroyed, but others will be created.

Lately we read a lot, or so I perceive it, that this in general is quite catastrophic. A bit of a climate of panic is being created that I think should be handled carefully. Many times we focus only on the most tragic side, but AI is a tool that has many good things if handled properly.

For example, in recent years we have seen its capacity for much more preventative medicine. It can help us in learning, we can be much more selective with our students and adapt their teaching, be able to predict livestock diseases, fight climate change, do more sustainable things or better manage your store’s stock.

There are many positive aspects that we must learn to take advantage of and protect ourselves from those that can harm us.

What dangers can there be with AI?

The dangers, for example, are that the system is behaving inappropriately and the person does not detect it, let’s say that their supervision is not as strict as it could be. But that is a human error from which we are not free even with AI.

It is also a profession that is still heavily biased towards the male gender and it is important that we become aware that part of the future is going to be designed with technology. How we get there, how we want the future to be is important, because the design of these tools requires us to be aware of biases and requires everyone’s participation.

But I think it’s helping people by empowering them in decision making. Imagine that you are a doctor and you have a lot of symptoms and doubts. You ask a colleague, in this case, an AI, and it narrows your chances of deciding. It helps you, but the final decision is up to you. Just like the algorithm on a platform can tell you what to watch, but ultimately you decide, you don’t let the machine decide for you.

It is true that AI is something in which we are making a lot of progress and that it is important to regulate.

Can it be regulated or is it like putting doors to the field? We have already seen what happens with the internet and the ‘deep web’, for example.

The European Union has been concerned about this for a long time. We are going slow, but there is a proposal.

Talk about this topic began in 2018 when the high-level artificial intelligence expert group was created that produced these guidelines for reliable artificial intelligence. There, there was already talk of a human supervision of AI and aspects such as sustainability, the absence of bias or security are looked at.

For example, human supervision is one of the basic points contemplated in European regulations, that is, that the application of any artificial intelligence system will always have a human supervisor throughout the entire operation initiation process, in the collection of the data, in the sectors behind its application.

In the EU we were pioneers and now we see that there is a great influence of companies, the majority from outside the EU, as in the United States, who insist on the need for this regulation.

It is something that must be done worldwide, we are working on it. The important thing is to take the first step.

Can everything be regulated? The answer is complex because AI is complex and it is clear that zero security does not exist here or anywhere. For example, we regulate and make traffic laws and that does not prevent accidents.

Global regulation would always be desirable, but it is difficult to do so. You just have to see the Kyoto protocol… Not all countries sign it and you have no way to force them to do so. Aside from the European Union, convincing the world’s other major AI centers, such as China and the United States, that regulation is necessary is not easy.

I think that outside of the media noise we should all be concerned because it is important to regulate this technology and constant monitoring of intelligent systems must be arbitrated.

Lately we have seen headlines and experts saying that AI can lead to the extinction of humanity… Is this so?

It’s hard to say how far artificial intelligence is going to go, but machines always have to have a way to interrupt them, to turn them off.

We are designing them by people… Just like people are working with nuclear energy. So I think it’s important to detect if there’s a problem and set security and application rules.

But the way I see it is that it happens with the AI ​​as it happened with the cars when they appeared. At first it was thought that they were going to be extremely dangerous, that they could kill people and that the speeds they reached could denature our proteins. Today we know that this is not the case and it is a technology that we have under our control, we have regulations in this regard, etc.

Can you surpass human intelligence and become aware?

Almost all AI systems exceed our intelligence, it just happens that in a certain field.

Most of the AIs we have are narrow niche: capable of having a very high level of intelligence in a very specific field. For example, the AlphaGo machine (which was taught to play go, a board game) can beat the world champion of go, but if you want it to play chess, you have to teach it, it can’t do it alone.

They may be great at diagnosing a type of cancer, but they are not good as general practitioners because the knowledge that is needed is broader.

And about consciousness… You can, in quotes, model it.

There are robots that can model feelings and can appear to have real awareness, but we don’t even know how certain processes of awareness occur in humans, so it’s very complex and vast.

Although we see tools such as chats, which seem more transversal because they are language-based, in reality what these machines do is be able to predict the next word in a text. He is a very sophisticated seeker, but he is not capable of deep reasoning because he is not aware. It’s like you have some kind of highly intelligent trained parrot.

Now you can receive notifications from BBC Mundo. Download the new version of our app and activate them so you don’t miss out on our best content.