Navigating privacy in the AI age

By Evans Sagomba | 11h ago

It is no secret that many AI systems, including popular platforms such as ChatGPT, have privacy invasive features turned on by default

In today’s digital age, where artificial intelligence seems to be woven into the fabric of everyday life, one question looms large: How do we protect our data when using these powerful systems? Have you ever wondered whether the friendly chatbot you use for advice is quietly storing every word you utter, or, worse yet, using your insights to train its algorithms without your explicit consent?

Privacy-invasive features

It is no secret that many AI systems, including popular platforms such as ChatGPT, have privacy invasive features turned on by default. When you fire up a chatbot, did you realise that your prompts, those seemingly trivial messages, might be collected and assimilated into the training data of ever evolving language models? Most users are blissfully unaware that behind the scenes, every input you share is potentially feeding the very model designed to interact with you.

Companies often hide these settings in layers of complex menus, leaving the average user, whether a tech savvy millennial in Harare or a street vendor in Mbare, without the necessary tools or knowledge to safeguard their privacy. We must ask ourselves: can we truly trust these systems that are designed to learn from our every interaction if they operate cloaked in obscurity?

Why opting out of training matters

Let’s consider the implications of having our everyday conversations used to train AI models. The data that fuels these sophisticated systems mostly comes from web scraping; an approach that collects content posted online without seeking explicit user consent. This means that when you share a personal anecdote on social media or confide something sensitive in an online forum, you could inadvertently be contributing to the data pool that trains AI technologies.

Even for those living in regions under strict data protection laws, such as the European Union’s General Data Protection Regulation, the option to opt out of AI training is becoming increasingly elusive. In recent news, Meta announced that data from EU-based users would continue to be used for training purposes, yet there was no clear mention of how users could disengage from this process. How can we protect our privacy when even the most regulated companies are hard to pin down on this issue?

For Zimbabweans who may not benefit from robust local data protection legislation, the call to act wisely is even more urgent. It is important to remain informed about the platforms you use and, wherever possible, check their policies on AI training. Always ask: Is my data being used the way I intend it to be?

Keep Reading

Turning off model training

If the idea of your inputs being used to train an ever-more powerful machine feels unsettling, you are not alone. Many are unaware that when you interact with systems such as ChatGPT, there is often an option to turn off model training. For example, within ChatGPT’s settings, you can visit the “data controls” area and disable the toggle that says “improve the model for everyone.”

By doing so, you actively choose not to contribute your data to the development of the system. But do most users know this option even exists?

This setting is hidden away in the digital labyrinth of user configurations, a reminder that protecting one’s privacy online often requires diving deeper than what meets the eye. So, next time you use an AI-powered tool, ask yourself: Am I comfortable with every interaction being stored and possibly used to refine the system further?

Erasing the digital footprint

Another privacy setting that has sparked significant debates in recent months is the “memory” feature.

Introduced by AI developers such as OpenAI and now adopted by others such as xAI, this feature allows chatbots to remember previous conversations in order to personalise future interactions.

While this may sound convenient, it raises a critical concern: what if all your personal details, your preferences, and even your mistakes are being stored indefinitely?

On ChatGPT, memory operates through “Saved Memories” and “Chat History.”

Given the potential risks ranging from data leakage to adversarial exploitation, the recommendation is clear: deactivate the memory feature whenever possible.

But, is this simple step enough in a world where many other privacy-invasive features are enabled by default?

Consider this: if we are to preserve our privacy in an era dominated by digital interactions, shouldn’t we be extra cautious about every piece of our personal history?

For many Zimbabweans, ensuring that your digital conversations remain transient and private is not just about safeguarding data; it is about protecting your integrity in an increasingly interconnected world.

Privacy-preserving behaviour

Even if you take the bold step of disabling model training and deactivating the memory feature, is that enough to secure your privacy? The short answer is: not entirely. Privacy-preserving behaviour also requires a shift in how we interact on digital platforms. It is not merely a matter of clicking settings; it is also about the information you choose to share.

You might ask: Isn’t it up to me to decide what to say? Certainly, but the reality is more complex. When using AI systems, every piece of data you input contributes to a vast repository that could potentially be accessed by unintended parties or exploited in future AI developments. It is crucial to exercise caution and restraint. Avoid sharing overly personal details or sensitive information that you would prefer not to become part of the broader data landscape. Have you ever paused to think about whether every personal detail should be shared with an algorithm?

In Zimbabwe, where data infrastructure is still developing and regulatory oversight may not be as strict as in other regions, the imperative for privacy-preserving behaviour is even more pronounced. Whether you are a business owner or simply an everyday citizen leveraging digital tools, understanding the nuances of what you share online can make all the difference in protecting your privacy.

New challenges on the horizon

Privacy is not a static concept; it evolves as technology itself advances. Today, we face emerging challenges that could reshape the landscape of online privacy.

One such trend is the increased use of advanced analytics and deep learning techniques that enable AI systems to reconstruct detailed profiles from seemingly innocuous data points.

Ask yourself: could that seemingly harmless conversation with your virtual assistant be pieced together with data from your web activity to form an exhaustive portrait of your habits and preferences?

Recent trends indicate that some companies are exploring even more intrusive methods.

For instance, there is growing concern over the use of adversarial techniques, where nearly imperceptible changes in data input could be exploited to extract sensitive information. Moreover, as AI systems become more sophisticated, the lines between data storage for functional performance and data usage for profitability blur further. In our quest for convenience, are we unwittingly signing away more of our private lives?

This is not a challenge faced only by technologists or privacy experts. It is a matter that affects every everyday person, from the small-business owner in Avondale to the market vendor in Gaza. The convergence of convenience and privacy invasions prompts us all to ask: How much of our personal history are we willing to trade for the sake of efficiency?

The future of privacy, UX and responsible AI design

Looking ahead, the future of privacy in AI will depend greatly on how companies design their user interfaces (UX) and the principles underpinning responsible AI design. Currently, many AI systems prioritise user engagement and rapid improvements over privacy transparency. This is why privacy settings are often hidden away, and invasive features remain enabled by default. How can we expect users to protect themselves if the very tools they interact with daily are designed in ways that obscure critical privacy options?

Imagine a future where every interaction with an AI system is accompanied by clear, easily accessible information on data usage and privacy settings. What if, rather than cloaking these features behind convoluted menus, companies adopted responsible AI design that places user privacy at the forefront? This future would allow users to make informed decisions with full clarity around how their data is processed, stored, and utilised.

For Zimbabweans and others around the globe, responsible AI design is not merely a luxury—it is a necessity. In a world where digital and physical lives are increasingly intertwined, transparent privacy controls could empower individuals to safeguard their personal information more effectively. It is high time for companies to ask themselves: Should not the protection of personal data be fundamental to the design of any interactive system?

A call to action: What can you do today?

In practical terms, there are several steps you can take right now to better manage your privacy when using AI systems like ChatGPT:

Check your settings: First and foremost, explore the settings of the AI platforms you use. Look for options related to data controls, such as toggles to disable “model training” or other features that store your inputs. Taking control of these settings is the first step towards reducing unwanted data contributions.

Disable memory features: If you are using chatbots with memory capabilities, disable them whenever possible. While personalised responses might be tempting, the privacy cost of saving your interaction history could be enormous.

Think before you share: It is a timeless piece of advice in the digital age: always consider whether the information you share online is something you would be comfortable with being stored indefinitely. Does your casual conversation with an AI contain details you would rather keep private?

Educate yourself: Stay informed about the latest developments in AI privacy. Knowledge is power, and understanding the ways companies use or misuse your data can help you make more informed choices about where and how you interact online.

Demand transparency: As ordinary users and particularly as citizens who deserve privacy, demand more clarity and transparency from companies. Whether through regulatory channels or by influencing public opinion, every voice counts when it comes to urging companies to prioritise responsible AI design.

Engaging in a wider dialogue

While technical settings and privacy toggles are important, the conversation about privacy in the AI era should not end at the individual level. Ask yourself: Isn’t it time that we, as a community, engage in a broader dialogue about data rights and digital sovereignty? In Zimbabwe, as in many places around the world, our voices need to be heard when it comes to digital privacy.

By participating in community discussions, attending digital literacy workshops, or even engaging with local media on the topic, you help to create an environment where informed choice becomes the norm. This is not only about protecting ourselves; it is about shaping the future of technology in a way that respects and honours our freedoms.

Responsible AI design: A shared vision for the future

The challenges we face today in preserving privacy within AI systems are symptoms of a deeper issue. At the heart of it lies a tension between the quest for technological advancement and the rights of individuals to control their data. Responsible AI design must therefore reconcile these two imperatives by ensuring that systems are built with privacy as a core tenet rather than an afterthought.

Imagine a world where an AI’s user interface clearly explains every data usage policy in plain language, where every setting is designed with transparency in mind, and where users are empowered to choose the level of personal data they wish to share. Is this not a future worth striving for? The responsibility is not solely on the companies who develop these systems—it also lies with us, the users, to demand innovation that is both intelligent and respectful of our privacy.

Real-world reflections from Zimbabwe

For an ordinary person in the streets of Harare, privacy concerns might seem distant compared to the immediate challenges of daily life, but in reality, these issues are intimately connected to our well-being. When our personal information is used without our permission, or when sensitive details are stored indefinitely in systems we barely understand, our trust in technology is eroded.

Take, for instance, the local entrepreneur who uses AI-powered tools to streamline business operations. While these tools boost productivity, they may also be sharing critical business strategies or personal financial data without explicit consent. In such scenarios, safeguarding privacy is not just a matter of personal choice; it is an economic imperative.

Furthermore, the youth, many of whom engage actively on social media, must come to terms with the reality that every post, tweet, or Facebook update could eventually be repurposed as training data for an AI system. In a context where information is power, should we not be cautious about what we share and whom we trust with our digital footprints?

Concluding thoughts: A collective responsibility

So, where do we stand today? In a world awash with data, the responsibility to protect our privacy falls on both individuals and technology companies alike. While it is heartening to know that options exist, such as deactivating model training and memory features, the onus remains on each one of us to act as vigilant custodians of our privacy.

As artificial intelligence continues to evolve, the design of these systems must evolve too. In future iterations, we hope to see interfaces that prioritise transparency and place data protection front and centre. But until that day arrives, every cautious click and every informed decision counts.

In the final analysis, it is not enough to rely solely on the goodwill of developers or the regulations imposed by distant policymakers. Instead, it is incumbent upon us, whether you are in a bustling urban centre or a small township, to educate, question, and even challenge the status quo. The digital revolution should empower us without compromising the sanctity of our personal lives.

So next time you interact with an AI system, consider the invisible strings that tie your data to faraway servers. Ask yourself: Am I in control, or am I just another data point in some vast machine-learning algorithm? And more importantly, what will you do to ensure that your digital life remains a private and secure extension of yourself?

Our privacy is not merely an abstract right, it is our tangible shield against unforeseen intrusion. It is time to demand clarity, embrace responsible AI design, and actively protect the intimate details that form the mosaic of our lives. After all, in the quest to harness the power of artificial intelligence, shouldn’t our fundamental human rights remain at the very core of technological development?

Let this be a call to citizens across Zimbabwe: Stay informed, stay vigilant, and never be afraid to ask the critical questions. Our digital future, much like our heritage, deserves the utmost respect and care.

As we stride into this brave new digital era, let us remember that while technology may connect us, it is our shared responsibility to ensure that connection does not come at the expense of our privacy. Whether you are a student studying at a local university, a business owner adjusting to digital innovations, or simply a curious citizen seeking to protect your personal space, the time to act is now. By making informed choices today, we lay the stepping stones for a future where technological progress and personal freedom walk hand in hand.

In this dialogue of progress and privacy, every question you ask and every precaution you take paves the way for a more secure and respectful digital landscape. Isn’t that a future worth fighting for?

By embracing both the convenience of AI and the right to privacy, Zimbabwe’s citizens and, indeed, people around the world, can shape a future that honours tradition, protects individual rights, and heralds responsible innovation. Let us remain steadfast and inquisitive, ensuring that every stride we take into the future is measured, mindful, and, above all, respectful of our personal space. In the end, the conversation about privacy will continue to evolve with every technological breakthrough. And as this dialogue deepens, we, the people, must be ever ready to question, learn, and adapt. After all, our privacy is not just a feature, it is a fundamental part of who we are.

This article is an invitation to dialogue and reflection, a gentle reminder that in the digital world, as in our daily lives, every prudent choice counts. Stay safe, remain curious, and protect what is rightfully yours.

Dr Sagomba is a doctor of philosophy, who specialises in AI, ethics and policy researcher, AI governance and policy consultant, ethics of war and peace research consultant, political philosophy and also a chartered marketer- [email protected]/ LinkedIn; @ Dr. Evans Sagomba (MSc Marketing)(FCIM )(MPhil) (PhD)/ X: @esagomba