In the year 2024, the primary focus of cybersecurity apprehensions can be encapsulated in the form of two letters – AI (or if one prefers a broader perspective, gen AI). Establishments are only scratching the surface when it comes to comprehending the risks and advantages associated with this technology. Despite the potential it holds to enhance data security, maintain compliance standards, and facilitate swift threat identification, malicious actors are utilizing AI to expedite their social manipulation schemes and compromise AI models using malware.
While AI garnered considerable attention in 2024, it wasn’t the solitary cyber threat burdening organizations. The issue of credential theft persists, evident by a 71% surge year-over-year in attacks employing compromised credentials. The shortage of skilled personnel remains a challenge, leading to an additional $1.76 million expenditure for companies following a data breach. With the rise in cloud dependency among businesses, the escalation in cloud breaches shouldn’t come as a surprise.
Despite the challenges, there have been positive advancements in cybersecurity over the preceding year. CISA’s Secure by Design initiative enrolled over 250 software vendors to enhance their cybersecurity practices. Additionally, CISA unveiled its Cyber Incident Reporting Portal to streamline the exchange of cyber-related information among organizations.
Whereas last year’s cybersecurity prognostications primarily revolved around AI and its implications on the future operations of security teams, this year’s forecasts also accentuate the significance of AI, underscoring the inseparability of security and AI in shaping the future, both positively and negatively.
Enclosed are this year’s anticipations.
Ubiquity of Shadow AI (Akiba Saeedi, Vice President, IBM Security Product Management)
Shadow AI is poised to manifest as more prevalent – and precarious – than previously perceived. Day by day, businesses are deploying an increasing number of generative AI models within their systems, at times without their explicit cognizance. In 2025, enterprises will come face to face with the scale of “shadow AI” – unsanctioned AI models utilized by personnel without proper oversight. Shadow AI poses a substantial jeopardy to data security, and corporations that effectively confront this issue in 2025 will employ a blend of unambiguous governance policies, comprehensive employee training, and vigilant detection and response mechanisms.
Metamorphosis of Identity (Wes Gyure, Executive Director, IBM Security Product Management)
The conception of identity within enterprises will continue to evolve in the aftermath of hybrid cloud and application modernization endeavors. Acknowledging that identity has emerged as the new security perimeter, enterprises will persist in transitioning towards an Identity-First approach, overseeing and safeguarding access to applications and crucial data, inclusive of gen AI models. In 2025, a pivotal element of this strategy will entail establishing a robust identity fabric, a platform-agnostic integrated suite of identity tools and services. When executed meticulously, this will serve as a welcome respite for security professionals, bringing order to the chaos and risks stemming from the proliferation of multicloud environments and disparate identity solutions.
Uncover cybersecurity services
Collaborative Efforts in Threat Management (Sam Hector, Global Strategy Leader, IBM Security)
Cybersecurity teams cannot operate in silos efficaciously any longer. The landscape of threats stemming from generative AI and hybrid cloud adoption is swiftly evolving. Concurrently, the threat quantum computing poses to contemporary standards of public-key encryption will become inescapable. With the evolution of novel quantum-safe cryptography standards, there will be a push to recognize encrypted assets and expedite the modernization of cryptography management. In the coming year, thriving organizations will be those wherein executives and diverse teams mutually craft and enforce cybersecurity strategies, ingraining security within the organizational ethos.
Preparation for Post-Quantum Cryptography Standards (Ray Harishankar, IBM Fellow, IBM Quantum Safe)
As organizations embark on the transition to post-quantum cryptography in the ensuing year, agility will be imperative to ensure systems are adept at embracing sustained transformation, particularly amidst the continual expansion of post-quantum cryptography standards by the U.S. National Institute of Standards and Technology (NIST). NIST’s inaugural post-quantum cryptography standards served as a clarion call to the world, signaling that the time has arrived to embark on the journey to achieve quantum-proof status. Equally crucial is the concept of crypto agility, assuring that systems can swiftly adapt to new cryptographic mechanisms and algorithms in response to evolving threats, technological advancements, and vulnerabilities. Ideally, automation will streamline and hasten this progression.
Data as a Fundamental Element of AI Security (Suja Viswesan, Vice President of Security Software Development, IBM)
Data and AI security will emerge as indispensable constituents of trustworthy AI. “Trustworthy AI” is commonly construed as AI that is transparent, equitable, and preserves privacy. These attributes are pivotal. However, if the AI and the underlying data lack security, all other aspects stand compromised. As businesses, governments, and individuals interact with AI more frequently and with greater stakes, the realms of data and AI security will be perceived as even more vital components of the trustworthy AI formula.
Organizational Learning on the Duality of AI’s Boons and Hazards (Mark Hughes, Global Managing Partner, Cybersecurity Services, IBM)
While AI transitions from a test concept to widespread deployment, enterprises will cherish the advantages of enhanced productivity and efficiency, encompassing the automation of security and compliance tasks to shield their data and assets. Yet, organizations must remain cognizant of the potential misuse of AI as a novel tool or conduit for threat actors to breach entrenched security protocols and procedures. Corporations ought to embrace security frameworks, adhere to best practice suggestions, and instate guardrails for AI promptly — to address both the opportunities and threats linked with the rapid progressions in AI.
Enhanced Comprehension of AI-Boosted versus AI-Driven Threats (Troy Bettencourt, Global Partner and Head of IBM X-Force)
Guard against AI-boosted threats; prepare for AI-propelled threats. A distinction exists between AI-driven and AI-assisted threats, dictating the approach organizations should adopt concerning their proactive security stance. Presently, AI-powered assaults, such as deepfake video scams, remain limited; the prevalent threats are mainly AI-assisted — wherein AI aids threat actors in fabricating variations of existing malware or crafting more sophisticated phishing email bait. To counter prevailing AI-assisted threats, organizations should prioritize implementing end-to-end security measures for their AI solutions, encompassing the protection of user interfaces, APIs, language models, and machine learning operations, while also planning defenses against potential AI-driven attacks in the future.
The prognostications impart a crystalline message: grasping how AI can be beneficial or detrimental to an organization is imperative for safeguarding your company and its assets in 2025 and beyond.