Information Literacy

The challenge is no longer finding information. It's deciding what deserves belief.

Information literacy has always involved more than finding information.

Libraries, educators and researchers have long recognised that locating a source is only the beginning.
We also need to evaluate its credibility, understand its context and decide whether the evidence justifies the conclusions being drawn.

What has changed over the years is the environment in which those decisions take place.

Information is more abundant than ever.

Search engines organise vast quantities of material.
Social platforms select and amplify content according to systems most users cannot inspect.
Artificial intelligence can generate convincing explanations, arguments and recommendations in seconds.

The difficulty is no longer just information access.
It's determining what deserves our attention, what warrants investigation and what we have sufficient reason to believe.

Information literacy is the capability to find, evaluate, interpret and responsibly use information to develop understanding and make informed decisions.

That definition remains relevant for life, whether reading an academic paper, evaluating a news report, investigating a workplace problem or questioning an AI-generated answer.

But those conditions have become more demanding over time.

We increasingly encounter information that has been selected for us, shaped around our preferences or generated without a transparent relationship to its underlying evidence.

The task is to exercise judgement in an environment where convincing information can reach us faster than we can properly evaluate it.

Information Is Abundant. Evidence Still Matters.

The cost of producing and distributing information has fallen dramatically.

Artificial intelligence can summarise research, explain unfamiliar concepts, translate documents and produce plausible answers almost instantly.

That capability is useful.
But it also creates a problem.

Producing a convincing explanation is not the same as establishing whether that explanation is true.

An AI-generated answer may accurately represent a substantial body of research. It may also combine reliable information with unsupported assumptions, omit important qualifications or attribute claims to sources that don't exist.

The presentation alone cannot tell us which is happening.

This is why the established disciplines of information literacy remain essential.

We still need to examine sources, compare evidence, recognise uncertainty and understand the relationship between a claim and the information supporting it.

The distinction between information and evidence is an important one.

Information tells us what is being claimed. Evidence gives us reason to accept, question or reject that claim.

Even reliable evidence requires interpretation.

A credible study may answer a narrower question than its headline suggests.
A technically accurate statistic may conceal important context.
An expert may speak confidently about a subject beyond their particular expertise.

Information abundance doesn't remove these problems.
It actually increases the number of claims competing for our judgement, while making some of them considerably easier to mistake for established knowledge.

Digital Information Literacy in an AI-Mediated World

Information literacy used to begin with a relatively simple problem.

Where can I find reliable information?

However, digital environments and generative tools have changed the question.

Now we also need to ask:

• Why am I seeing this information?
• Who benefits from my attention?
• What has been removed from the context?
• And what has been generated rather than discovered?

Search engines rank.
Social platforms personalise.
Recommendation systems predict what might keep us engaged.
Artificial intelligence can generate a seemingly authoritative answer without making clear where its claims came from.

The information environment is no longer just something we draw on as needed, but something which is increasingly shaped around our lives.

That changes what digital information literacy requires of us.

In a saturated, ambiguous information environment, it's not enough to know how to build searches effectively.
We also now need to understand how information is selected, framed, amplified and generated.

A person can be highly skilled at finding information and still be highly vulnerable to manipulation.
They can know how to use a search engine without understanding why one result appears above another.
They can produce an excellent prompt without recognising whether the resulting answer should not be trusted.

Digital information literacy therefore sits at the intersection of information, technology and judgement.

This also helps clarify how several related literacies fit together.

• Information literacy concerns finding, evaluating, interpreting and using information responsibly.
• Digital literacy includes the capabilities required to navigate, understand and participate effectively in digital environments.
• Media literacy examines how messages are constructed, distributed and interpreted, including the influence of representation, ownership and persuasion.
• AI literacy adds an understanding of what artificial intelligence systems can do, how their outputs are produced, where their limitations lie and how to use them responsibly.

These disciplines overlap considerably. Nor are they competing explanations of the same problem.

A person evaluating an AI-generated news summary may need all four.
They need to understand the information being presented.
The digital system delivering it.
The media context shaping its interpretation.
The capabilities and limitations of the AI involved.

What connects these disciplines is the need to exercise judgement rather than just operate the technology.

The goal isn't to somehow become suspicious of everything.
Rather, it's to become more deliberate about what earns our attention, our trust and what warrants further investigation or verification.

That's increasingly the real literacy, and challenge of the digital age.

Why Misinformation Is Difficult to Recognise

Misinformation is often treated as a problem that can be solved by just giving people vetted or more accurate information.

Sometimes that helps. But it doesn't fully explain why misleading claims remain persuasive even when good contradictory evidence is available.

People don't encounter information as neutral observers.
We interpret claims through existing knowledge, experience, identity, expectations and emotional responses.

A claim may feel credible because it confirms something we already suspect.
A dramatic warning may capture our attention because it concerns something we fear.
A confident explanation may seem authoritative because it removes uncomfortable uncertainty.

None of these reactions establishes whether the claim is true.

Digital platforms complicate the problem by algorithmically influencing which information reaches us and how often we encounter it.

Repeated exposure can make a claim feel familiar. That familiarity can be mistaken for credibility, particularly when we cannot easily recall where the claim originated.

Generative AI introduces another difficulty where information can be presented in a fluent, coherent and seemingly authoritative form without a reliable evidentiary foundation.

The challenge is therefore partly informational and partly cognitive.
We need to evaluate the claim itself while remaining aware of the assumptions and reactions influencing our interpretation.

Information literacy asks us to examine both the evidence before us and the judgement we're bringing to it.

The Belief Filter for Information Literacy

Not everything that captures our attention deserves our belief.
Yet information can influence our thinking long before we've consciously decided whether to trust it.

We notice a headline.
Encounter a confident explanation.
Hear a claim repeated by several people.
Read an AI-generated answer that seems to make sense.

By the time we begin evaluating the claim, it may already have influenced how we understand the problem.

The Belief Filter is a framework for making that process more deliberate.

It follows four connected stages:

Information → Attention → Interpretation → Judgement

1. Information - What is actually being claimed?

Before evaluating a statement, establish what it says.

Separate the specific claim from its presentation, emotional framing or implied conclusion.

A statement such as “AI improves student learning by 40%” sounds precise, but what does improvement actually mean? Which students? What was measured? Compared with what?

If we cannot identify the claim clearly, we cannot evaluate it properly.

2. Attention - Why has this captured my attention?

Information competes for limited attention.
Some claims attract us because they're surprising or alarming.
Others confirm our existing beliefs or appear to offer a simple explanation for something complicated.

Digital systems also influence what we notice through ranking, recommendation and repetition.

This stage asks us to recognise why a particular claim has become prominent in our thinking.

Attention is necessary for evaluation, but it's not evidence of credibility.

3. Interpretation - What meaning am I assigning to it?

We interpret information through existing knowledge, experience and assumptions.

Those mental models help us understand unfamiliar material.
They can also cause us to overlook alternative explanations or read more into a claim than the evidence supports.

Ask what context might be missing, what assumptions connect the claim to its apparent conclusion and whether another interpretation is plausible.

4. Judgement - What does the evidence justify?

Only after identifying the claim, examining our attention and questioning our interpretation do we reach the central evaluative task.

What should we believe?

Five questions help:

• Source. Where did the claim originate, and can we identify its original source?
• Evidence. What reliable evidence supports it?
• Context. What information might change its meaning?
• Uncertainty. What remains unknown, contested or insufficiently supported?
• Revision. What evidence would cause us to change our conclusion?

These questions don't guarantee certainty, but they do help us determine how much confidence the available evidence warrants.

Four Possible Responses

Evaluation doesn't always end with a simple choice between believing and rejecting a claim.

We may:

• Accept provisionally. The evidence is sufficiently strong for the purpose at hand, while remaining open to revision.
• Reject. The claim is contradicted by reliable evidence or lacks support where support is reasonably required.
• Suspend judgement. The available evidence doesn't justify a confident conclusion.
• Investigate further. The claim matters enough to warrant additional verification before we rely on it.

These responses can overlap. We may suspend judgement while investigating further, or accept a claim provisionally while continuing to examine its limitations.

The purpose of the Belief Filter isn't to make every decision slower.
But to prevent attention, familiarity or apparent confidence from being mistaken for evidence.

Judgement determines what deserves belief, how strongly we should believe it and when we should remain uncertain.

How to Evaluate AI-Generated Information

Consider a university lecturer preparing a proposal to introduce generative AI into their course.

They ask an AI assistant to summarise research on whether AI improves student learning.

The AI assistant responds with a confident claim:
“Research shows that students using generative AI achieve 40% better learning outcomes.”

The statement sounds useful. It offers a precise figure, supports a plausible argument and appears relevant to an important decision.

But should the lecturer believe it?

This is a hypothetical example, not a finding from an identified study. Our purpose is to demonstrate how the Belief Filter works.

1. Information - Identify the claim

The lecturer first needs to establish what the statement actually means.

Does 40% refer to higher examination scores, faster assignment completion, improved retention or a different measure?
Was the improvement observed in one subject, across several institutions or in a controlled experiment?

Without those details, the claim is too ambiguous to support a meaningful conclusion.

2. Attention - Recognise why it is persuasive

The statistic appears precise.
It also supports a proposal the lecturer may already favour.

That combination is important. A numerical claim can feel authoritative, particularly when it confirms an existing expectation or bias.

The lecturer needs to recognise that the claim's usefulness to their argument doesn't establish its accuracy.

3. Interpretation - Examine the assumptions

Suppose the underlying research measured how quickly students completed an assignment using AI.

That might demonstrate improved task efficiency.
It would not necessarily demonstrate improved learning.

Students could produce stronger immediate outputs without developing a corresponding improvement in independent understanding or judgement.

The distinction between performance and capability changes how the evidence should be interpreted.

4. Judgement - Examine the evidence

The lecturer now investigates the claim.

Can the original research be located?
Does it contain the reported figure?
What outcomes were measured?
Were there appropriate comparisons?
What limitations did the researchers acknowledge?

If the AI assistant cannot identify a verifiable source, the lecturer has insufficient grounds to repeat the statistic as an established finding.

Even if a genuine study exists, its conclusions must still be interpreted within the limits of its methods and context. AI is not the only information source that can be wrong.

The Outcome

The lecturer may conclude that the original claim should be rejected, that the evidence supports a narrower proposition or that further investigation is necessary.

They can still explore the potential benefits of AI in education.
But the proposal should rest on evidence that actually supports the argument being made.

This is what responsible information literacy looks like in practice.

The aim isn't to identify whether every sentence was written by a machine.
But to recognise when a claim deserves confidence, when it requires qualification, and when we should withhold belief.

A plausible answer is a starting point for investigation, not a substitute for it.

Information Literacy and Critical Thinking

Critical thinking is central to information literacy.
It allows us to examine assumptions, evaluate arguments, compare evidence and revise conclusions when better information becomes available.

But the relationship between the two deserves some precision.

Critical thinking concerns how we reason about claims and problems.

Information literacy applies those reasoning capabilities to the discovery, evaluation, interpretation and responsible use of information.
It also requires an understanding of how information is produced, organised, distributed and made accessible.

These capabilities can reinforce one another.

A person may be able to identify logical weaknesses in an argument but struggle to locate the evidence needed to evaluate it.
Another may be highly skilled at finding credible sources but less capable of recognising when those sources have been interpreted incorrectly.

Neither capability is sufficient on its own.

From Finding Sources to Understanding Disagreement

Consider a student researching a contested question.

They locate five peer-reviewed articles.
All five are relevant.
All come from credible academic publications.
Yet their conclusions differ.

The student's task isn't just to decide which article is most authoritative.
It's to understand why those findings differ.

Perhaps the researchers studied different populations, used different methods, measured different outcomes or worked from different assumptions.

The disagreement may reveal genuine uncertainty rather than poor research.

A student who can locate five relevant articles has demonstrated information retrieval.
A student who can explain why those articles disagree has begun demonstrating the evaluative judgement that information literacy requires.

As AI makes retrieval and summarisation easier, that distinction becomes increasingly consequential.

The enduring educational task is to develop people who can examine evidence rather than just collect convincing explanations.

Institutions Are Becoming Judgement Infrastructure

Libraries, universities, schools, news organisations and professional bodies have long helped people access, evaluate and use information.

Artificial intelligence makes that responsibility more consequential, not less.

When information was difficult to obtain, access infrastructure played a particularly visible role.
Libraries organised collections.
Universities provided scholarly resources.
Publishers and professional organisations helped people discover specialised knowledge.

Those functions remain necessary.

But access alone cannot resolve the challenges of an information environment in which plausible claims can be generated, personalised and distributed at extraordinary scale.
Institutions also need to strengthen what we might call judgement infrastructure.

From Access Infrastructure to Judgement Infrastructure

Access infrastructure helps people discover, retrieve and obtain information.

Judgement infrastructure helps people evaluate evidence, understand context, recognise uncertainty and use information responsibly.

These aren't competing responsibilities.
They are complementary pathways in helping people develop an informed understanding.

A university library, for example, may provide access to scholarly research while also teaching students how to evaluate competing sources and recognise misleading interpretations.
An academic may help students move beyond collecting evidence towards understanding what that evidence can reasonably establish.
A professional association may provide guidance that helps practitioners distinguish established knowledge from emerging claims.

In each case, the institution contributes more than information alone.
It creates the conditions in which better judgement becomes possible.

What Changes in the Age of AI?

As AI becomes more capable of retrieving, summarising and generating information, institutions need to examine what they are helping people learn.

• Can students recognise when an AI-generated reference is fabricated?
• Can professionals distinguish a plausible recommendation from an evidenced conclusion?
• Can people identify when an automated summary has omitted qualifications that materially change the meaning?
• Can they remain intellectually independent when a system consistently produces answers that appear more confident than the underlying evidence warrants?

These are increasingly important questions for education, professional practice and wider public life.

The opportunity isn't in institutions competing with AI at producing information.
It's to help their people develop the judgement required to use increasingly capable information systems responsibly.

Information Literacy Within Stoic Futurism

Within Stoic Futurism, information literacy belongs to the broader development of human judgement.

It concerns a question that becomes increasingly important as information grows more abundant:

What deserves belief?

Human Judgement extends that inquiry into deciding what deserves weight and action when information is incomplete, priorities compete and consequences matter.

Decision-Making examines how those judgements become choices.

Systems Thinking helps us understand the environments, incentives and feedback mechanisms that influence which information reaches us and how we interpret it.

These capabilities are all connected.

A person cannot exercise sound judgement without some understanding of the evidence informing their decisions.

But we must remember that accurate information alone doesn't guarantee wise action.

Information literacy helps establish what we have reason to believe.
Judgement determines what we should do with that understanding.

What Deserves Belief?

The goal of information literacy has never been to just accumulate information.
It's always been about developing an understanding that can withstand scrutiny and remain open to revision.

Artificial intelligence has made that responsibility much more visible.

We can now generate convincing explanations faster than ever before.
We can also encounter claims selected and presented in ways that influence us before we've consciously evaluated them.

The tempting response might be indiscriminate scepticism. But that's its own dead end.
Rather, the response should be disciplined judgement.

We need to recognise what captures our attention, examine the assumptions shaping our interpretation, and decide how much confidence the evidence actually deserves.

Sometimes that means accepting a claim.
Sometimes it means rejecting one.
And sometimes the most responsible conclusion is that we don't yet know.

Information literacy doesn't tell us what to believe. It helps us determine what deserves belief.