Anyone who uses the internet has encountered artificial intelligence. A robot that calls and offers to listen to information about a new credit card. A social media feed that suggests interesting videos. A chatbot that answers standard questions in a live casino. All of this is AI.
But this technology is used in different areas of life. Neural networks help solve ecological problems. For example, scientists from UT Austin used them to find an enzyme that breaks down plastic.
Algorithms detect abnormalities in X-rays, analyze CT and MRI data. This speeds up the diagnosis of many diseases, including cancer. The most famous virtual diagnostician is IBM Watson.
Programs drive cars. The program compares sensor data with information from its database to select the optimal speed and trajectory for the car. Tesla uses this technology.
“Smart” machines take care of safety. For example, banking programs detect suspicious transactions and block cards. And other software detects forest fires and alerts rescuers.
Computers become tutors. The US Department of Defense created an educational platform. It adjusts lectures to each student’s abilities to make the lessons as beneficial as possible.
Programs can recognize faces and fingerprints. In China, it’s not necessary to take out a card to pay for goods. It’s enough to look at the validator. A special sensor reads facial features, verifies them with a database, and the bank deducts the money from the person’s card.
AI-controlled drones deliver goods. Amazon has been developing this idea since 2013. In some African countries, drones deliver water, food, and medicine.
Chatbots write texts and code. ChatGPT writes poems, scripts, and posts for social media. It can create code, find bugs, and debug the process.
Risks and Problems Strong and super-strong AI still remain something out of the realm of science fiction. There is no talk of the infamous machine rebellion yet, but it’s not without problems.
Resource-intensive. To create and train an algorithm, you need a team of developers, large amounts of data, and computational power, as well as large financial investments. The costs of maintaining them are in the millions of dollars. Such expenses are only affordable for large corporations.
Inaccuracy. Machine learning requires a lot of data. Gaps in them affect the final result: the program can make mistakes.
In 2020, during a football match, the program confused the ball with the bald head of the referee. The camera followed the referee’s movements and ignored what was happening on the field. The developers explained that the algorithm got confused because there were too few bald people in the database. Another example: algorithms identified a green hill as a sheep. It was all because sheep are often photographed against such a background.
Narrow specification. A person can use their knowledge in different fields. Programs are far from this yet. They specialize only in a specific task and cannot apply their skills to another. A neural network cannot simultaneously predict the demand for bananas and solve math problems.
Development prospects Precedence Research predicts that by 2030, the market volume will reach $1.6 trillion. It’s three times more than in 2023. Experts attribute this growth to the boom in internet technology. There are currently no prerequisites for the industry to start declining. But, business giants are investing hundreds of millions of dollars in application development.
At the same time, a machine uprising is definitely not looming in the next ten years. The power of an artificial “brain” depends on the complexity of the neural network. Here, the human brain is still out of reach.
