ML & Data ScienceAI Trends in 2022/2023 (Updated)

Introduction

Machine learning, neural networks, and all related technologies continue to evolve. AI has penetrated into many areas of life and work. According to Sundar Pichai, CEO of Google Inc. and parent company Alphabet, the impact of artificial intelligence (its weak form) on humanity is comparable to the impact that the use of fire or the beginning of the use of electricity had.

The influence of AI technologies on humanity is gradually growing, and it will continue to increase next year. Experts predict several trends that will become the main factors in the development of AI technologies in the second half of 2022 - to the first of 2023.

Additional labor force

The point is that AI technologies will significantly increase the capabilities of conventional tools used in companies. Machine learning provides an opportunity to improve the efficiency of people's work. For example, in medicine, the method of analyzing X-rays using AI is becoming increasingly used. Machines detect signs of diseases in the early stages, which a person in many cases is unable to do. A lot of pictures pass through an ordinary doctor a day and it is clear that over time concentration and attention fall, so the doctor may miss some important moments.

Predictive analytics is actively used in the industry, a technology that makes it possible to determine the probability of failure of the unit or its element even before something actually broke. Accordingly, the continuity of production, and hence business processes, improves.

Next year, the technology will continue to develop, and experts from different industries will increasingly use tools that make it possible to sort huge amounts of information needed to solve various tasks.

Speech recognition

Speech recognition technologies have moved far ahead compared to what it was just a couple of years ago. So, now machines are able to simulate human speech very realistically - it is extremely difficult to recognize where a person speaks and where a machine speaks.

Not so long ago, the release of GPT-3 from OpenAI, the most advanced (and the largest) the language model ever created, consisting of about 175 billion "parameters" - variables and data points that machines can use to process language. Now OpenAI is working on a successor to GPT-4, which will be even more spectacular. Although the details have not been confirmed, according to some estimates, the model will contain up to 100 trillion parameters, making it 500 times larger than the GPT-3. Theoretically, this is an opportunity to model speech that is indistinguishable from human speech. 

AI in information security

This year, the World Economic Forum called cybercrime a potentially more dangerous factor than terrorism. As machines penetrate into all spheres of our life and work, as mentioned above, hacking and cybercrime are inevitably becoming an increasingly urgent problem.

Literally, every connected device that is added to the network inevitably becomes a potentially vulnerable element that an attacker can use against a company or an individual. As the networks of connected devices become more complex, identifying vulnerabilities also becomes more complicated.

And that's where AI comes to the rescue. By analyzing network traffic and learning to recognize patterns suggesting malicious intentions, smart algorithms help network security experts. Some of the most important AI advances we will see in 2023 are likely to be in this area.

Low- and zero-code in business

The shortage of qualified developers has now reached unprecedented heights. The global deficit, according to estimates of analytical agencies, is several tens of percent - up to 40% in some industries.

In many cases, business, science, and the state cannot do without qualified developers. But in some moments, you can use an alternative - no-code systems that allow you to develop services and applications for ordinary users. So far, the quality of such systems leaves much to be desired, but they are improving.

And here artificial intelligence plays a very important role, which is able to translate various user actions into program code. By the way, speech recognition also finds its place in this niche. Natural language recognition technologies and modeling in the near future may make it possible to use only voice commands or interface interaction for writing services and programs.

Of course, there is still a long way to go, but the zero-code reality is already somewhere nearby.

Autopiloting

Another industry that continues to actively develop. More recently, it was impossible to imagine a car that rolls itself along public roads without the participation of a person. In science fiction, of course, a lot has been said about this, but in real-life cases have only recently begun to appear.

By the way, according to Elon Musk, the number of accidents after the penetration of technology piloting in the automotive industry may decrease by up to 50%. Tesla is making the most significant progress in this direction so far. But competitors don't sleep either. Thus, the companies Waymo (created by Google), Apple, GM, and Ford, and a number of Chinese startups have made very far progress in the issue of creating autopilot systems.

And this applies not only to cars but also to other types of transport. So, next year, the world's first ship that crosses the Atlantic under the control of artificial intelligence may appear. It is currently being developed by IBM in partnership with the non-profit organization ProMare.

AI in Art

We are talking about creating works of art with the help of AI technologies. Writing articles and short stories, creating musical compositions, paintings and even video games is most likely not a very distant reality. Now the first attempts are being made in this direction, and in many cases they are successful.

Probably in 2023, we will hear about new breakthrough projects that have allowed machines to master new areas for themselves. More recently, it was believed that the computer is not available for the game of go, because it seems to be based more on intuition than on strict rules. But no, artificial intelligence has long defeated man in go, many computer games.

Well, now it's the turn of AI to create games, texts, videos, and paintings.

Conclusion

In general, these are the main trends that can be expected in the development of AI technologies next year. If you have any comments on this topic, let's discuss it in the comments below here on #MRBlog and on our Medium publication page.