Ahead of the Global Media and Information Literacy Week, running from24 to 31 October 2026, we speak with Dr Theresa Josephine Seipp, Postdoctoral Researcher at the AI, Media & Democracy Lab and the Institute for Information Law at the University of Amsterdam.

---

How has generative AI changed the way we should think about media literacy and online disinformation?

Theresa Josephine Seipp: Generative AI enables the creation of various types of disinformation. For instance, it makes it easier to produce more convincing visual and deepfake videos. This means fake content or disinformation can look more realistic, making it much harder for people to tell what's real and what's not, so media literacy becomes crucial for everyone.

It is not just about identifying whether something looks or sounds authentic, but about understanding the sources behind videos or posts, their content, and potential intentions. We need to rethink media literacy in this context. Additionally, addressing disinformation involves understanding the role of journalism and considering AI literacy, including how algorithms, engagement metrics, and incentives shape what is visible online.

---

Does age matter when it comes to vulnerability to disinformation in the AI era? Are younger or older people more at risk, and in different ways?

Theresa Josephine Seipp: When it comes to lifelong learning and literacy, I believe age matters. However, age does not inherently make people more or less vulnerable to AI. For example, younger people tend to understand social media, algorithms, and related areas better. Older individuals generally have less awareness of these aspects. However, younger people mostly consume news via social media, which exposes them to highly personalised content.

Therefore, every age group faces distinct challenges with disinformation and AI, making it important to address these issues across all demographics.

---

What are the most important AI literacy skills people need today to recognise and resist disinformation?

Theresa Josephine Seipp: I'm not an absolute expert on AI literacy skills, but it's obviously more important than ever to verify sources and cross-check with credible reporting, which again stresses the importance of quality journalism and the need for trust in journalism.

AI literacy is crucial not just for the general public but especially for journalists, because maintaining trust in accurate journalism requires journalists to verify information. If journalists make mistakes, such as copying AI-generated text without fact-checking, using false sources, or quoting inaccurately, it can undermine public trust in journalism. Addressing AI literacy is thus vital, as disinformation isn't always a straightforward lie; deepfakes are different, but disinformation can be more subtle.

I also think it's important for people to understand that misinformation or misleading information can nudge people towards a certain view, a more radicalised perspective, or reinforce biases. These aren't necessarily outright lies, but they still carry an intention behind them. So, an important part of AI literacy is understanding that it's not merely about labeling something as right or wrong.

Maybe a last point is that people should be more aware of how algorithms actually work, what reposting or resharing means for engagement, and how this can lead to false content being spread throughout social media or even amplified further. It is also crucial to consider what kind of AI chatbots you're using and where you're uploading your data. I think that's also part of AI literacy, which many people still lack.

---

Are we putting too much responsibility on individuals to spot false or misleading information? What should platforms, AI companies and regulators be doing instead?

Theresa Josephine Seipp: Everyone has a responsibility, but expecting individuals to verify everything online is unrealistic; it's impossible for us all to do that given the overwhelming amount of information we encounter daily. Therefore, a governance system involving multiple actors is necessary.

Policymakers and regulators play a crucial role in supporting AI media literacy and ensuring AI platforms aren't the only ones moderating content through their own tools. We all know that not all of these platforms are owned by entities that have democratic values and may have questionable ties to political interests and aren't truly neutral; they don't remove content impartially. So, you wouldn't want to give them too much political influence or control over what content we see, and they shouldn't be the ultimate arbiters of truth.

Existing transparency obligations for AI-generated content do not necessarily offer a solution either. Many studies have shown that people are often confused by overused transparency labels. These labels, which indicate only that content is AI-generated, do not necessarily reveal whether the content is true. Instead, they can be misleading and add to the confusion.

A crucial area where regulators must take positive action is in cases like deepfake pornography, disinformation, or violations of individuals' rights through deepfakes. This responsibility cannot be left solely to platforms or individuals; regulators have a clear obligation to intervene and regulate effectively.

---

Are current EU rules, such as the AI Act, Digital Services Act and European Media Freedom Act, doing enough to address these risks and power imbalances?

Theresa Josephine Seipp: They are all important steps. We all know that they address small pieces of a big puzzle, as I mentioned before. But on their own, they're not enough. Also, as I mentioned, there's a lot of focus on transparency obligations, which, in terms of disinformation, won't do much, because they won't tell people whether a specific piece of AI-generated content is accurate, trustworthy, or manipulated. So all of these transparency obligations are already not very effective.

None of them truly tackle media organisations' reliance on big tech and AI firms, which poses a problem because without independent media, there is no credible counterpower or source of alternative information. While the current EU laws include important aspects, none of the legal provisions is effective unless properly enforced.

A significant issue is the slow enforcement of the EU Digital Markets Act (DMA), especially against major U.S. tech firms. We've all seen Vance and Trump threaten the EU if it enforces certain rules on these U.S. companies, as they want to maintain their monopoly. So that brings the debate from regulatory concerns to broader geopolitics and security policies. So, clearly, more decisive action is needed.

---

So should AI literacy also include understanding who owns AI systems, how they make money, and who has power over the information we see?

Theresa Josephine Seipp: Yes, 100%. That's what I was just pointing out: the issue of independence, which is crucial in understanding who owns these AI systems, how they generate revenue, and the extent of their concentrated power.

It's also important to recognise their connections with political figures. Many people mistakenly see AI companies as purely innovative and commercially driven, assuming they lack political interests. However, for example, Palantir - the American software company specialised in collecting, combining, analysing, and securing large amounts of data, receives funds from the U.S. Immigration and Customs Enforcement (ICE), supporting its surveillance system, and OpenAI has made large donations to Trump. This information is vital for the public to know.

Achieving this understanding requires investigative and independent journalism with the resources to operate without financial or technological dependencies on these companies. To effectively scrutinise their influence, we need a truly independent media as a counterbalance to concentrated power and disinformation.

Related article