The Question of Accountability: Informed User or Responsible Systems?
The User: always the “victim”
In the current era, artificial intelligence seems to intersect with most aspects of our lives, and I have observed a subtle yet consistent pattern of victim-blaming. While some may find the term heavy and insensitive when applied to digital media users, let me explain why it fits within the current discourse.
From personal experience, people keep telling me to use AI critically. However, without concrete examples or practice, this advice lands as a vague buzzword, leaving me overwhelmed as I try to build a skill that only develops over time. Equally unhelpful is the fact that big tech repeatedly develops technology designed to dim critical thinking. By building artificial intelligence tools optimised for speed and convenience, they inadvertently reduce the very friction that critical thinking depends on, encouraging what is known as cognitive offloading. Society places responsibility on digital media users, which shifts focus from the root cause of the problem: opaque platforms. These systems conceal their inner workings by design. In response, the European Commission calls for large online platforms to start holding “more responsibility and less opacity.”
The complexity of teaching critical thinking
Even so, AI literacies urge users to critically assess the “accuracy, explicability, and fairness of AI systems”, meanwhile also resisting “algorithmic bias” and the “erosion of autonomy caused by opaque platforms.” However, even if users meet all these conditions, they may still feel a sense of agentic hubris over AI, thinking they have mastered how to “outthink” it, which is exactly what the act of critical usage tries to resist.
Are Gen Z even more responsible?
Particularly since the imagined “ideal citizen” is expected to navigate disinformation and configure complex privacy settings, effectively clocking into a job every time they access digital media. Yet, much of this use is passive by design, as the rise of doom-scrolling shows. Infinite entertainment makes continuing easier than putting the phone down. Labelled ‘digital natives’, Gen Z is assumed to be fluent in these environments, yet also the most deeply locked into them, as social, educational, and professional life increasingly runs through a handful of platforms. Consequently, consent becomes coercive: if refusing a platform leads to exclusion, then accepting its terms is less a choice than a condition of participation. Asking users to take responsibility for ‘informed use’ is unrealistic, and disregards the structures that make opting out so costly.
This postures the tension between negative and positive freedom. Negative freedom is the absence of interference: these platforms are not forcing users’ hands to sign up to their digital spaces. Positive freedom is the genuine capacity to act on a choice, which involves having alternatives and understanding of what is being chosen. Users hold the first, while missing the second, as their agency exists on paper but, in practice, platforms hollow it out.
The Attention Economy shows how platforms erode this freedom. Human attention is the product; platforms commodify and exploit it at the expense of the user. The reality is platforms profit from emotionally charged content, hence why engagement-driven algorithms work so well. This is in part due to the mass “Platform Surveillance”, in which users’ engagement is tracked to further bolster it. A logical impasse confronts users, who face a choice that exists on paper but offers little room to manoeuvre in practice. Thereby, digital users cannot strictly bear responsibility; responsibility should track freedom. Platforms design systems to trigger people and boost engagement and mindless interaction. Therefore, we must rethink the push for critical thinking to ensure users develop the ability to recognise when platforms are playing them. In gaining this awareness, they can take a step back, reevaluate, and pause before reacting.
Why AI literacy is not enough
Critical AI literacies should not be disregarded; in fact, they hold the key. It is the approach that must face scrutiny. The focus spans intellectual scaffolding as a critical defence to empower digital media users. The way I see it, critical thinking and assessment of accuracy, explicability, and fairness come from remaining grounded. Nonetheless, nobody should use this set of capabilities, or the ability to execute them perfectly, as a scapegoat that undermines the deeper structural problems faced.
Instead, we can view humility and emotional regulation alongside larger systems’ accountability and adopt them as enriching tools. Not to be mistaken as an additional duty that exacerbates users’ culpability. Humility involves recognising the scale and sublimity of AI systems, and the limits of what any individual can verify alone. Meanwhile, emotional regulation implies fighting back against the “rage-bait” that larger platforms give rise to. By allowing our feelings to be felt, but not letting them influence the way in which we continue to approach media. These skills should be regarded as complementary, in unison with platforms rearing transparency and trust.
By fostering a culture of inquiry, users can begin developing and building on competences to resist the allure of falsehoods. For example, when enraged by a politically charged video that seems like an agenda plant, taking a step back to question: What was the intent behind its publication? Is it even true? This allows us to engage critically but also assess the algorithmic bias and AI systems’ fairness. Despite this, user resilience can only go so far when systems are built against us, working to market off our attention and engagement. We must devise systems that do not require extraordinary heedfulness from ordinary users, so they can enjoy them safely.
Taming the Giants
With all this being said, “the informed user” prototype frames a “deficit” in judgement as a personal failure rather than a failure of the system. What gets obscured is the structural technological dependency on hyperscaler cloud platforms and large language model ecosystems. Small states like Malta, and its inhabitants, are structurally captive to these larger infrastructures that they neither own nor govern. Further emphasising the irony of expecting digital media users to be informed about systems entirely foreign to them.
The obvious question would be: why are these platforms so difficult to rein in? Mostly because the EU negotiates from a weak position. It has no frontier model of its own, and regulation is slow across all 27 states. Malta’s technological dependency and the user’s trust in opaque platforms are two sides of the same coin: one at the state level and the latter at the individual level. Generally, both issues can only be resolved through acknowledging that sovereignty, in any form, will remain stagnant as long as its underlying structural dependencies are dismissed.
Fortunately, the use of DigComp 3.0 aids in identifying where critical judgement breaks down, so that policy can respond, as opposed to treating the breakdown as a personal failure. Its role is not to produce the exemplary “informed user”, but to build an environment where the user will ideally not need to be hyper-vigilant.
Shifting the burden of integrity
The fight begins by holding systems accountable to a high ethos of public digital infrastructure and open standards that value and protect user agency. By way of this, the burden of integrity should rightfully shift to the architecture, not its consumers. This is not to say that the act of victim-blaming must be replaced with victimising ourselves as media users. With anything, we must critically examine the information we consume, keeping in mind that “if the product is free, you are the product”. Victimising assumes we have no power to conduct our own actions going further.
The problem is not an absence of agency, but the conditions it operates in. The goal is to expand users’ positive freedom and focus on building modular resilience, remaining open but not powerless. By creating systems that are transparent and trustworthy, the user is unburdened of unrealistic expectations. With the influx of information and constant mental stimulation users are faced with in their daily lives, it is not sensible to assume they will tediously read through terms and conditions and think about the algorithmic bias each and every time they scroll on Instagram or TikTok. Especially a generation brought up on television and short-form, high-stimulation content.
The importance of the control variable
Not every individual acts, thinks, and proceeds the same way, further highlighting the individual variation and diversity in our society. To put it this way, instead of focusing on confounding variables, perhaps the attention should be garnered on controlling variables. If regulators held systems to ethical standards and principles, their usage would not generate so much confusion and, consequently, cryptic terms of use.
Undoubtedly, a user must be cautious and engage critically with the media they consume. However, users should not feel culpable for lacking expert technological knowledge of algorithms which are infiltrating their everyday lives. AI literacy is highly valuable but must not overshadow the expectation of systems to create ethical digital environments.
