The year 2024 will bring widespread adoption of AI, and will also be a turning point for the field of cybersecurity due to the introduction of new data encryption algorithms. RBC Trends tells us what awaits the IT sector in 2024
Analysts predict that in the coming year, we should expect a wide spread of applications based on artificial intelligence, the heyday of the "Internet of things" for robots and the development of computing as a service.
The rise of AI and the battle for chips
Technology consulting firm GP Bullhound predicts that 2024 will be the year of artificial intelligence and technologies based on it will begin to play a central role in shaping the future of the technology industry. Analysts have identified several key trends in the field of AI in 2024.
Large language models will become smaller. Such AI models or agents are already being developed by the developer community. It is assumed that small models will become no less productive and at the same time will have more accurate knowledge in the chosen field.
A boom in the acquisition of training data will begin. So, OpenAI has already launched a program to purchase datasets from partners to train their AI models. They will probably be included in the training set of the GPT-5 model.
Analysts at Air Street Capital believe that at least one AI company will launch an IPO in 2024.
At the same time, it is predicted that next year a large-scale competition for processors for AI training will begin. Sam Altman, the head of OpenAI, has already stated that the ChatGPT subscription was temporarily suspended due to the heavy load on the company's computing power. Analysts at Air Street Capital predict that in 2024 financial institutions will create so-called GPU debt funds, which will become an alternative to venture capital in financing computing. In addition, in their opinion, due to the increase in workloads, a large artificial intelligence company (for example, OpenAI) may buy out an artificial intelligence chip manufacturing company.
Widespread adoption of AI applications
According to forecasts from the Institute of Electrical and Electronics Engineers (IEEE), artificial intelligence-based applications will be used in a wide variety of fields as early as 2024. AI will help optimize data, perform complex tasks, and make decisions with human precision. IEEE names several potential applications for applications with embedded AI:
real-time detection of cybersecurity vulnerabilities and prevention of attacks;
improving the efficiency of supply chain automation and warehouse operations;
assistance and acceleration of software development;
customer service automation;
speeding up the screening of applicants, vacancies, recruitment and hiring of staff;
accelerating disease research and drug discovery;
automation and stabilization of power supply sources.
Analysts at Air Street Capital agree that self-learning AI agents will be able to perform complex tasks in 2024 — develop games, solve problems in science, etc.
Forrester adds that companies will develop their own "brand language models" that will be trained on audience data for various marketing mechanisms. So, they will be able to produce creative content personalized for a specific person using fonts, colors, message tone and other branding elements.
The growth of peripheral computing
Peripheral computing is a process aimed at localizing data warehouses that allow more efficient collection of data from devices that produce information, as well as from users who consume it.
Gartner analysts note that peripheral data centers help companies cope with the growing volume of data, especially when it comes to working with AI systems. By 2024, innovations have significantly reduced the cost of developing and deploying peripheral systems, and the technologies themselves have reached maturity.
In the coming year, this will allow companies to focus on implementing more advanced encryption algorithms, secure boot mechanisms and hardware security features, as well as take full advantage of peripheral and cloud computing. According to Accenture analysts, almost 50% of the surveyed companies use this comprehensive approach, and 79% plan to fully integrate peripherals with the cloud in the next three years.
In addition, in 2024, with the wider adoption of 5G, new options for using peripheral computing will appear. Companies will be able to interact with data regardless of whether they are located in a centralized on-premises, public or private cloud data center. This will ensure greater connectivity of the Internet of Things in various fields, will allow for more efficient implementation of automation and integration of AI tools. McKinsey predicts that 70% of companies will use hybrid or multi-cloud technologies, management tools and processes.
Finally, peripheral computing as a service (EAAs) will be developed in 2024, when companies will be able to choose a specific subscription to scale their peripheral computing resources without significant infrastructure investments.
Analysts at StartUs Insights, in connection with the development of peripheral computing, point to the possibility of widespread adoption of the Internet of robotic things. It provides detection, monitoring and tracking of robots that perform certain tasks. Connected robots collect and send data to peripheral computing platforms, receiving feedback from them in almost real time. For example, the Portuguese startup Sky Powerlines is developing drones based on the Internet of Things for aerial photogrammetry. It captures images of power lines to create 3D maps, and then the SaaS platform analyzes them and offers analytical reports. This allows enterprises to better plan power lines and spend less money on their inspections.
Hyperautomatization
Forrester analysts predict that in 2024, leading companies will transform automation processes, taking into account the potential of artificial intelligence. Thus, solutions based on large language models will act as autonomous assistants or digital colleagues, as well as be used in those workflows where efficiency is required. For example, AI-based chatbots will be able to provide customers with round-the-clock support and other services.
As predicted in Forrester, this is what it will lead to:
10% of all operations will be performed by digital employees — these can be sales, delivery, adaptation or product support operations, as well as administration, maintenance and support services;
autonomous delivery will reach a new level — most of the work will be performed by unmanned loaders, delivery robots and drones;
Enterprise application vendors will cover 35% of all automation costs, and they will focus on developing visualization and process analytics tools that will be embedded in enterprise application platforms.
It is expected that by 2024, 80% of IT companies from the global top 600 ranking will begin to implement hyperautomatization processes in their roadmaps.
Post-quantum Cryptography
As expected, 2024 will be the beginning of the widespread introduction of post-quantum encryption algorithms. These algorithms will help protect data from potential attacks using quantum computing. NIST (National Institute of Standards and Technology of the USA) announced that in 2024 the first three such algorithms will be ready for mass use.
IT companies are already taking the first steps in this direction. So, Google began to implement quantum-resistant encryption algorithms in Chrome version 116. In addition, the company offered an open version of the FIDO encryption protection algorithm, which implements secure login to sites without passwords. Google expressed the hope that this innovation will be supported by major web browsers. Signal Messenger has also updated the end-to-end communication encryption protocol to include a quantum-stable algorithm.
According to experts, companies and organizations need to start implementing post-quantum encryption without waiting for such attacks to become possible. This will protect the data not only from threats in the future, but also from current attacks, the authors of which use the tactic of "collect now, decrypt later".
Comentários