#TECH: Building safe AI environments

New Straits Times

25-04-2025

MALAYSIAN enterprises are increasingly engaging in conversations about the technology's potential and pitfalls.
Yet, despite high confidence levels in its adoption, a significant gap remains in actual readiness and long-term planning.
According to Kyndryl managing director (Malaysia and Indonesia operations) Effendi Azmi Hashim, artificial intelligence (AI) is now a fixture in boardroom discussions.
"Every time I meet clients, including top executives and boards, the conversation inevitably circles back to AI.
"Interestingly, even seasoned board members with decades of experience are still asking: what is AI?"
Effendi believes the question isn't one of ignorance, but a request for clarity amid the technology's rapid evolution.
To address this, Kyndryl developed the AI Readiness Report, based on a global survey of 3,200 senior executives.
The report offers a unique lens into the attitudes and preparedness of enterprises around the world.
One standout finding is that 86 per cent of global leaders surveyed expressed confidence in their AI implementation efforts.
However, only 30 per cent felt their AI systems were future-proof — a significant gap that reflects the speed at which AI technologies are advancing.
"We're seeing AI evolve so fast. Last year it was all about ChatGPT, now it's DeepSeek.
"That kind of rapid change challenges companies to ensure their systems, governance and security can keep up."
FOMO-DRIVEN ADOPTION
One of the key issues identified by Effendi is the fear of missing out (Fomo) — a trend driving many organisations to adopt AI tools without a clear strategy.
"I've seen companies invest in tools like Copilot or Gemini, but they're not even using them.
"They bought in because everyone else was doing it, not because they had a defined purpose."
This hasty adoption often leads to poorly integrated solutions, underutilisation, or worse — security risks stemming from public tool usage.
Effendi shared a telling example where some firms were drafting sensitive corporate reports using public tools like ChatGPT, despite having access to enterprise-grade alternatives such as Microsoft Copilot.
"That's a classic case of poor change management. Same tool purpose, but drastically different implications in terms of data security and governance."
AI CONFIDENCE, BUT MISSING FOUNDATIONS
While Malaysian firms are eager to adopt AI, Effendi said many lack the fundamental infrastructure and cultural preparedness to make that adoption sustainable and secure.
"When companies come to us, we often find a disconnect. They want to launch AI projects, but their current systems — from security to data pipelines — are not ready."
This is where tools like Kyndryl Bridge come into play. The platform offers real-time observability into an organisation's IT posture, helping identify gaps in performance, consumption and integration readiness.
Security remains a major concern, especially among Malaysian boards and CEOs. Effendi said local leaders are generally more aware of AI's potential to expose vulnerabilities.
"They don't just want to know what AI is. They want to know how they can be protected — today and in the future."
THE HUMAN FACTOR
Beyond technology, the human element is a critical factor in successful AI deployment.
The AI Readiness Report identifies culture and trust as key ingredients, both of which require proactive change management.
"You have compliance officers, but do you have a responsible AI manager?" Effendi asked.
"Who owns AI in the company? Is it the CIO? HR? That debate is still ongoing."
Effendi said AI should be approached like any organisational transformation — with leadership, communication, training, and buy-in across all levels. "Without this, even the best technologies may fall flat.
"Younger employees are ready — they're already using AI tools. But others are cautious or even resistant. That mindset gap has to be addressed."
BUILDING SAFE AI ENVIRONMENTS
To help organisations experiment with AI in a secure way, Effendi suggested companies "play" with the technology first before going all in.
He recommended timeboxed pilot projects that allow companies to experiment with specific AI workloads while managing expectations and reducing risk.
LOOKING AHEAD
Despite the hype and the hurdles, Effendi remains a strong advocate for AI — but only when deployed responsibly and with purpose.
"AI offers incredible opportunities for optimisation, cost savings and innovation.
"But without clarity of purpose, you won't know what value you're aiming to achieve."