Watched by design: How AI is quietly reshaping privacy landscape

INTRODUCTION:

In an age where artificial intelligence is increasingly embedded in everything from search engines and recommendation tools to cybersecurity platforms, privacy risks are no longer confined to high-profile data breaches. Instead, they manifest in quiet, persistent ways – through the invisible layers of AI that observe, infer, and adapt based on our behaviour. These systems don’t just collect what we say – they learn from how we move, what we hesitate on, and where we linger.

This talk sheds light on how AI is quietly transforming the privacy landscape – not through breaches, high-profile misuse or users’ carelessness, but through subtle design choices and background systems that constantly monitor our behaviour often without our explicit consent or understanding.

As AI becomes deeply integrated into the tools and services we rely on, the line between helpful functionality and intrusive surveillance becomes increasingly blurred. This session invites participants to ask: how much do we really know about what’s tracking us – and what can we do to take back control?

SESSION GOALS:

• Raise awareness of how AI-driven systems collect and use behavioural data to shape user experiences and decisions.

• Bridge the gap between technical realities and public understanding of AI’s role in privacy risks.

• Empower attendees to critically assess AI systems and advocate for transparent, privacy-conscious design.

PRESENTATION CONTENT DESCRIPTION:

Part 1. Overview of how AI is changing the privacy landscape.

How AI is altering the meaning of privacy – not through dramatic data breaches, but through quiet, continuous observation.

*“Let’s talk Metadata”*: While most users focus on content, it’s behavioural data — scrolls, pauses, and clicks — that reveals the most. We’ll explore how AI systems use this “digital exhaust” to build detailed behavioural profiles.

*Invisible AI*: “But I’m not using ChatGPT, why should I care?” Even when users think they’re not engaging with AI, they often are – indirectly. This section covers how AI functions in the background of systems like websites and cybersecurity tools, shaping experience without disclosure.

*What’s New in AI That Harms Privacy?*: From emotion-aware assistants to expanded predictive analytics, we’ll highlight features that heighten privacy risks — and prompt deeper trust than users may realise.

Part 2. Understanding AI privacy concerns.

*Case studies*: 1-2 real-life examples that highlight how AI-driven data inference has resulted in unintended or unethical outcomes, including over-personalisation, manipulation, or exclusion

*Theoretical perspective*: 1-2 theoretical findings that highlight the power asymmetry between users and AI-driven systems (Privacy Calculus Theory, Parasocial Relationship Theory)

Part 3. Privacy risks mitigations

This part covers both user-empowering strategies and structural changes that reduce AI-related privacy risks: technical options, design strategies, organisational measures

Part 4. Can AI help fight against AI?

What if AI doesn’t have to be the problem? How can we turn it into part of the solution? Finally, we consider how AI itself can be used responsibly — to detect hidden tracking, enforce privacy settings, and increase system accountability.