Top IT trends of the coming year: how AI and automation markets will transform in 2024
AI will need more computing power
Artificial intelligence has become the main theme of 2023. As we move from one version of neural networks to another, the results of their work have noticeably improved, and this trend will continue in 2024. The main blocker to this development is becoming the technical limitations of companies. Sam Altman, the head of OpenAI, has already announced a temporary suspension of ChatGPT subscriptions due to a significant utilization of the company's computing power.
Air Street Capital analysts anticipate that financial institutions will create dedicated GPU (graphics processing unit) funds in 2024, which will become an alternative to venture capital in financing computing tasks. In addition, they foresee that due to increasing workloads, artificial intelligence companies (such as OpenAI) may acquire a chip manufacturer to strengthen their technological capabilities in this area.
Another factor affecting the pace of neural network development is the availability of data to train them. GP Bullhound analysts predict a growing demand for such data. For example, OpenAI has already launched a program to purchase datasets from its partners to train its artificial models. In all likelihood, this data will be included in the training set for the GPT-5 model.
Another expected trend is the specialization of neural networks. It is believed that the size of large language models will shrink. Smaller models will be just as efficient and yet have more accurate domain specific knowledge.
More AI applications for hyper-automation
It is predicted that in 2024, leading companies will transform automation processes, moving to so-called hyper-automation. The Institute of Electrical and Electronics Engineers (IEEE) predicts that by 2024, AI-based applications will be widely used in a variety of fields. AI will provide assistance in optimizing data, performing complex tasks, and making decisions with a level of accuracy comparable to humans. IEEE highlights several potential application areas for AI-enabled applications:
-
Identify real-time cybersecurity vulnerabilities and prevent attacks.
-
Insurance Risk Assessment.
-
Improving the efficiency of supply chain automation and warehouse management.
-
Support and accelerate software development.
-
Customer Service Automation.
-
Accelerate candidate screening, vacancy, recruitment and hiring processes.
-
Accelerating disease research and drug development.
-
Automation and stabilization of power supply sources.
Forrester predicts that by the end of 2024, 10% of all operations will be performed by digital workers, including sales, delivery, product support, administration and technical support services. At the same time, the use of AI for professional tasks reduces the trust of end consumers: for example, Gartner suggests that in 2024, up to 50% of content consumers will limit the use of social networks due to distrust of potentially generated content. The creation of so-called acoustic brands - companies that build their identity on their refusal to use AI - is also predicted. Building a balanced AI strategy and properly distributing tasks between neural networks and live employees is a challenge that companies from all industries will face in the coming year.
The rise of edge computing
Peripheral computing aims to localize data stores to more efficiently collect information from the devices that create data and from the users who use that data. According to Accenture analysts, nearly 50% of companies surveyed are taking this holistic approach, and 79% plan to fully integrate edge devices with the cloud in the next three years. In 2024, edge computing will evolve as a service (EaaS), allowing companies to choose a specific subscription to scale their edge computing resources without major infrastructure investments.
StartUs Insights analysts note that the development of edge computing opens up significant opportunities for the adoption of the "Internet of Robotic Things." This approach allows robots performing tasks to be discovered, monitored and tracked. Connected robots collect data and send it to edge computing platforms, receiving near real-time feedback. For example, Sky Powerlines, a startup from Portugal, is developing Internet of Things-based drones for aerial photogrammetry. They capture images of power lines, creating 3D maps that are then analyzed and provide analytics reports via a SaaS platform. This helps businesses plan power lines more efficiently and reduce their inspection costs.
Development of post-quantum cryptography
Post-quantum cryptography - cryptographic algorithms that are resistant to cyberattacks using quantum computers. These algorithms will protect data from potential attacks using quantum computer computations. NIST (US National Institute of Standards and Technology) announced that in 2024 the first three such algorithms will be ready for mass use and post-quantum encryption will become the new security standard.
In 2023, IT companies have already started to implement the first solutions based on this technology. For example, Google has started implementing quantum-resistant encryption algorithms in version 116 of the Chrome browser. In addition, the company presented an open version of the FIDO security algorithm, which allows you to securely enter sites without using passwords. Google hopes that major web browsers will support this innovation. Messenger Signal also updated its end-to-end communication encryption protocol by integrating a quantum-resistant algorithm.
Experts note the importance for companies and organizations to start using post-quantum encryption in advance, not waiting until such types of attacks become relevant. This will help protect data not only from future threats, but also from current attacks, where attackers can use "collect now, decrypt later" tactics.