Season 4 - Episode 34: Using AI with Confidence: How Security Matters
- PIMM Sthlm
- 10 hours ago
- 4 min read
AI is no longer only something we ask questions.
It is becoming something we connect to our documents, systems, workflows and decisions. That makes AI more useful, but it also changes the security conversation.
Because once AI gets access, the question is not only what it can do. It is what it can see, what it can access and what it can act on.
In this episode of Simply Briefed, Kristine Lium speaks with Steven Antoniou, Cybersecurity Solutions Specialist at Lenovo, about cybersecurity in the AI era, and why security should not be seen as the opposite of innovation.
Done well, cybersecurity can give people and organisations the confidence to use AI safely, responsibly and creatively.
Key topics
Why AI is moving from a traditional tool to something more active
Why access is becoming one of the most important security questions
Why policies are not enough without visibility and control
How shadow AI and embedded AI are already changing organisations
Why the right guardrails can help people use AI with more confidence
Episode 34: Using AI with Confidence: How Security Matters
– With Steven Antoniou, Cybersecurity Solutions Specialist at Lenovo
This week’s episode
As AI becomes more connected to the tools, data and workflows around us, cybersecurity becomes a foundation for confident adoption. This episode explores how organisations can create the structures that allow people to use AI safely without slowing down innovation.
For many organisations, AI adoption comes with a difficult tension.
On one side, there is the fear of risk. Sensitive data, unclear access, unknown tools, hidden usage and the possibility that one mistake could create a much larger problem.
On the other side, there is the pressure to move forward. AI is already part of everyday work, and people want to use it to save time, create better outcomes and solve problems in new ways.
The answer is not to close every door.
It is to understand which doors should be open, who should be able to walk through them and when we need to keep watch.
AI is becoming more than a tool
A central shift in the episode is the idea that AI is no longer just a passive tool.
When AI is used to answer a question, the interaction is fairly limited. But when AI is connected to documents, systems, databases, workflows and agents, it becomes part of how work gets done.
That makes it more powerful. It also makes the security question more complex.
If AI can access information, take action, interact with other tools or operate inside a workflow, organisations need to understand what it is allowed to do and where the limits should be.
This is where cybersecurity becomes less about saying no, and more about creating the conditions for safe use.
Policies are not enough
Many organisations try to handle AI use in one of two ways.
They either block everything, making it difficult for people to experiment and learn, or they allow almost everything and rely on a policy to guide behaviour.
But a policy alone does not create control.
If an organisation cannot see what tools are being used, what data is being shared or where AI is already embedded, it becomes difficult to know whether the policy is actually being followed.
Visibility matters.
So does the ability to act when something crosses a boundary. Not to punish people for trying, but to make the safe path easier to follow.
Shadow AI is already here
One of the challenges discussed in the episode is that AI is not always introduced through a formal decision.
Sometimes it appears inside tools that organisations already use. A new feature. A beta version. A button that suddenly offers to write, summarise or automate something.
This means AI can enter workflows before leaders, security teams or employees have fully understood what has changed.
That is why the conversation around AI security cannot only be about the obvious tools. It also needs to include embedded AI, shadow AI and the everyday ways people are already experimenting.
Confidence enables creativity
Security is often framed as something that slows innovation down.
But the episode offers a different perspective.
When people are afraid of making one mistake, they may stop experimenting altogether. They avoid the tool, even when it could help them do better work.
The right guardrails can change that.
If people know that sensitive information is less likely to be exposed, that risky behaviour can be identified and stopped when needed, and that there are clear structures around what is allowed, they can use AI with more confidence.
That confidence matters.
Because safe AI adoption is not only about preventing mistakes. It is about helping people explore, create and move forward without unnecessary fear.
Making the safe path the easy path
The goal is not to make AI harder to use.
It is to make responsible AI use easier.
That means understanding what is happening, deciding what should be allowed and designing structures that support the behaviour organisations want to see.
It also means moving away from the idea that cybersecurity and innovation sit on opposite sides.
In the AI era, they are increasingly connected.
Because the organisations that can use AI with confidence will not be the ones that ignore security. They will be the ones that make security part of how people are able to use AI well.
🎧 Listen now on Spotify!
Host: Kristine Lium
About the guest
Name: Steven Antoniou
Title: Cybersecurity Solutions Specialist at Lenovo
Background: With over 20 years of experience in IT and telecommunications, I work at the intersection of cybersecurity, endpoint management, and emerging AI governance. In my role at Lenovo, I support enterprises in gaining visibility and control over their device estate, strengthening security posture, and aligning with evolving regulatory frameworks such as NIS2 and the EU AI Act.
My focus areas include AI security and governance, endpoint and network protection, risk and compliance, and digital experience monitoring.
👉 Listen to the full episode of Simply Briefed to explore how organisations can create the structures, visibility and guardrails needed to use AI with confidence.
Comments