Back in 2018, during the South by Southwest conference in Austin, Texas, USA, the founder of innovative high-tech companies Tesla and SpaceX, Ilon Musk, said his prophetic and eccentric thesis that artificial intelligence (AI) is more dangerous than nuclear weapons. Is it really so?
Elon Musk's announcement, as a catalyst, has emboldened proponents of the concept of universal artificial intelligence, who predict a "revolt of machines" that will acquire abilities hitherto unknown to human natural intelligence and then destroy humanity.
The reality, however, is somewhat different.
Table of Contents
The limitation of artificial intelligence
We observe that current artificial intelligence technologies, while fantastically successful, have significant and fundamental limitations:
- Artificial intelligence systems successfully solve only a single type of problem - the one for which these systems were designed from the very beginning;
- They don't have the ability to "switch context" like humans are able to do;
- to fulfil their task, artificial intelligence systems require a certain amount of time for training, as well as information about a "reference truth". That is, a certain amount of truthful data is needed on which to train the system before putting it into operation;
- in moments of significant changes in external conditions, it becomes necessary to temporarily decommission the system for re-training under new conditions.
But AI in Moldova is impossible without people - those who define goals, organise training, generate data, define rules and algorithms for applying the systems, and put them into operation.
Risks of AI application in Moldova
We are all well aware that the uncontrolled use of artificial intelligence technology poses very real dangers to humans in particular:
- loss of jobs due to automation of routine repetitive operations;
- breach of confidentiality;
- Deepfakes is an AI-based video or audio synthesis method that can compromise anyone;
- Automated weapons that can attack objects without the intervention of human operators;
- erroneously biased solutions of the systems due to curvature of the initial training data.
- preservation of socio-economic inequality between different segments of the population;
- The creation of total control, in which a small group of people armed with the tools of artificial intelligence systems are empowered to regulate the lives of the majority of people through AI projects.
By understanding how modern AI systems are built and trained, it is clear that the risks can become real, not because of the systems themselves, but because of the people who design and implement them.
AI opportunities for Moldova
Globally, the concept of AI development is a great leap forward. It will help integrate innovative technologies into many spheres of human life:
- People will acquire personal robot assistants. AI can perform household functions: switch on lights, ventilation system, make a schedule, inform about traffic jams on the road, or recommend the nearest place based on preferences.
- Facial recognition technology to replace credit cards. In the near future it will be possible to pay with your "face" in shops instead of cards. Thus, there will be no need to stand in queues.
- Chatbots will become even more popular and advanced. Yes, chatbots are still "raw" today, but nevertheless, they are taking over the role of operators and consultants in online shops. And most importantly, they don't need to be paid a salary.
- Flats and houses will be "smart". Internet networks are already connecting hundreds of thousands of devices throughout Moldova. This applies not only to mobile phones, tablets and PCs, but also to other home gadgets: hoovers, TVs, air conditioners, heaters, etc. The so-called "Internet of Things" is gaining popularity - this is when devices work in conjunction with the Internet and among themselves.
- AI will become an author and composer. Artificial intelligence will be able to write books and music based on people's preferences.
- The use of virtual reality will become an everyday phenomenon. Many people tend to associate VR glasses with video entertainment, but their scope of application is potentially much broader: for example, in medical schools or sports.
Pros and cons of implementing AI
Despite all the pros and cons, the fact is that AI technologies are useful in terms of:
- Error avoidance. With proper programming, errors can be reduced to zero. For example, robotic surgical systems that perform complex procedures with incredible precision.
- Reducing risks to humans. Robots with AI can perform dangerous tasks in place of humans, such as defusing explosives.
- High labour capacity. The AI does not need rest and lunch breaks, it never gets sick and is always awake.
- Business Assistants. Many companies and online shops use AI for chatbots.
- Lack of empathy. Artificial intelligence is devoid of emotions, it is rational and practical, which can be used during the recruitment of personnel in the company, based solely on the professional skills of the applicant.
- Ability to do routine work. Checking documents for errors, sending out letters, etc. A person with the help of AI can get rid of routine and concentrate on creative activities.
But along with all the chips, there are downsides to implementing AI:
- The cost of software and hardware is high;
- it's impossible to teach an AI to think outside the box;
- some professions are being abolished and unemployment is rising;
- people become more lazy because of the reduced load on the brain's neurons;
- minimises banal human interaction;
- AI is good at routine tasks - no more and no less (it just doesn't know how to do better).
Despite the promising potential of artificial intelligence, it has divided scientists into two camps: some are convinced of its benefits, while others warn of its ability to destroy humanity. What do you think?
If you have found a spelling error, please, notify us by selecting that text and pressing Ctrl+Enter.
