AI Predictions for 2025: Insights from a Cyber Security Expert

AI adoption is accelerating, with significant growth in generative AI platforms and energy demands. Innovations in energy efficiency, like compute-in-memory architectures, are critical for managing increased consumption.

As artificial intelligence (AI) continues to evolve at an unprecedented pace, the year 2025 is set to bring transformative changes to the cyber security landscape. This was the focus of insights shared by Dan Karpati, Vice President of AI Technologies at Check Point Software. Karpati’s analysis highlights the challenges and opportunities AI presents across sectors such as energy, software development, and regulatory frameworks. Below, we outline the key developments and their implications for the industry.

AI adoption: accelerating growth and implications

The adoption of AI technologies continues to expand at an extraordinary pace. Generative AI platforms like ChatGPT, Claude, Gemini, and others are seeing widespread integration, with over 92% of Fortune 500 companies leveraging generative AI by the end of 2024. The global large language model market is projected to grow to $140.8 billion by 2033, according to Dimension Market Research.

However, this growth has environmental implications, including significant demands on energy, water, and land resources. With the number of data centres doubling over the past decade, Deloitte forecasts that AI workloads will represent 19% of total data centre power consumption by 2028. Addressing these challenges will require innovations in sustainable energy and advanced computational efficiency.

Energy demands and efficiency innovations

The surge in AI adoption has placed immense strain on energy resources. By 2034, energy consumption from data centres is expected to rise to 1,580 TWh annually, a figure comparable to India’s total energy use. To mitigate these pressures, advancements such as compute-in-memory (CIM) architectures, which reduce energy-intensive data transfers, are expected to play a key role. Additionally, nuclear energy and alternative power sources will likely become critical to meeting AI’s growing energy needs.

Responsible AI: A new era of software development

AI-driven software development is transitioning from code completion tools to full-scale code generation platforms such as CursorAI and Replit.com. While this advancement enhances productivity, it also introduces new risks, including the potential misuse of AI for creating sophisticated malware.

In response, the concept of responsible AI is gaining momentum. Security vendors and developers are collaborating to embed robust guardrails into AI models, ensuring they cannot be easily weaponised. Testing AI systems for vulnerabilities before they can be exploited will be a key focus area in 2025, requiring partnerships between software developers and cyber security experts.

Multi-agent AI systems: Defining the future of cyber security

The emergence of multi-agent AI systems will be a defining feature of 2025. These autonomous systems operate with minimal human intervention, enabling enhanced decision-making, threat detection, and real-time collaboration across networks. While attackers are expected to use these systems for coordinated cyber-attacks, defenders will employ them to share threat intelligence and neutralise threats more effectively.

This collaborative capability marks a significant evolution in how AI is used for both offensive and defensive cyber operations.

Ethical and regulatory developments in AI

As AI becomes more integrated into everyday operations, ethical and regulatory considerations are taking centre stage. The EU’s AI Act and the United States’ AI regulatory frameworks will introduce significant oversight, beginning in early 2025. These measures will require organisations to demonstrate transparency, reliability, and ethical considerations in AI deployments.

Industry-specific AI assurance frameworks are also set to emerge, helping enterprises validate the security and fairness of AI systems. These initiatives aim to mitigate risks, foster trust, and ensure compliance with new regulations.

Conclusion

The advancements in AI technology in 2025 present both opportunities and challenges. From energy efficiency innovations to responsible AI practices, collaboration between governments, private enterprises, and security vendors will be crucial to navigating this rapidly changing landscape. By embracing transparency, sustainability, and ethical practices, stakeholders can ensure that AI strengthens, rather than undermines, digital trust.

Cyber security professionals are urged to remain vigilant, adapting strategies to leverage AI’s potential while addressing its inherent risks.