Who Are You Listening To?

There has probably never been this much psychological information available to an ordinary person.

A question about why I keep procrastinating can take me, within twenty minutes, through a psychologist’s Instagram post, three people describing ADHD on Reddit, a YouTube video about the nervous system, an article quoting a research study, an AI assistant explaining executive functioning, and an influencer telling me that procrastination is a trauma response.

Some of these may be useful. Several may even be true.

They are not, however, the same kind of information.

The useful question is less Can I trust mental-health information online? and more: What kind of information am I looking at, what can this source reasonably tell me, and how much weight should I give it?

Here is the map I use.

A person describing their own experience

Reddit, personal blogs, comments sections and social media can give access to something professional literature often cannot: what an experience feels like from the inside.

If I want to know what starting therapy felt like for different people, what someone found difficult after an ADHD diagnosis, or how people experience grief years later, hundreds of first-person accounts can enlarge my understanding.

What they establish is much narrower.

One person’s experience cannot tell me how common that experience is, what caused it, whether I have the same condition, or what treatment I need.

Use lived experience for possibility and perspective, not diagnosis or probability.

The mental-health influencer

“Influencer” describes a medium more than a qualification. The person making the reel may be a psychiatrist, psychologist, therapist, coach, researcher or someone with no professional training at all.

So I would look past the follower count and biography first.

What are they trained in? Are they speaking within that area? Do they distinguish education from diagnosis? When discussing research, can I find what they are referring to?

Short-form content is particularly good at introducing a concept. It is much worse at preserving complexity. Attachment theory, narcissism, trauma, autism or dissociation may be useful places to begin reading after a three-minute video. Three minutes is rarely enough to decide that a concept explains a person.

Use influencers for discovery. Follow the interesting idea further.

The podcast or long-form interview

Longer formats allow context that reels cannot. You can hear someone develop an argument, explain qualifications, acknowledge exceptions and disagree with another expert.

But length isn’t evidence.

A confident person can be confidently wrong for three hours.

When something matters, separate this person gave a compelling explanation from there is good evidence for this explanation.

Use long-form content for understanding an argument. Verify consequential claims elsewhere.

Articles from hospitals, universities and professional organisations

These are often useful when I want basic, reasonably established information: What are common symptoms? What treatments exist? What does a particular diagnosis mean? When should someone seek professional help?

They tend to sacrifice some complexity for accessibility, but that is often exactly what introductory information needs to do.

Use them for orientation and established basics.

Research papers

A research paper gets closer to the evidence, but “a study found” does not mean “science has established.”

Studies vary enormously in quality. Samples may be small. Findings may not replicate. One population may not resemble another. Researchers themselves may interpret the same evidence differently.

Most of us also aren’t trained to critically evaluate every methodology we encounter.

For important questions, reviews, meta-analyses, consensus statements and clinical guidelines can often tell us more than one exciting new study.

Use research to investigate claims, not to collect studies that confirm what you already believe.

AI assistants

AI occupies a peculiar position because it can sound like all the sources above at once.

It can explain terminology, compare ideas, translate technical language into ordinary language, help formulate questions for a doctor or therapist, summarise material you provide, suggest search terms, or help you understand where different psychological theories disagree.

That makes it a very useful thinking and navigation tool.

But fluency can be mistaken for authority.

An AI system can misunderstand context, give outdated or incorrect information, invent details, flatten disagreement between experts, or produce a remarkably persuasive explanation from insufficient information. It also cannot perform the examination, observation, longitudinal assessment or clinical responsibility involved in actually treating someone.

So use AI to ask better questions, understand information and find what needs verification. The higher the stakes—medication, diagnosis, immediate safety, treatment decisions—the less sensible it is to let an AI-generated answer stand on its own.

Your therapist or another clinician

A professional brings something the internet generally cannot: information considered in the context of a particular person over time.

That doesn’t make clinicians infallible. Training differs. Theoretical orientations differ. Professionals have blind spots and limits to their competence. They can also be wrong.

What changes is the kind of claim they are able to make. A therapist who has worked with me for a year has information that a reel called Seven Signs You Have an Avoidant Attachment Style simply doesn’t.

Use professionals when the question requires assessment, individual context, treatment or responsibility for a clinical decision. And when something doesn’t make sense, ask how they reached their conclusion.

Whatever the source, ask a few more questions

When I encounter a psychological claim online, I find these useful:

Who is making the claim?

What kind of knowledge could this source reasonably have?

Are they describing an experience, offering an interpretation, reporting research or making a clinical claim?

Can I trace the information to somewhere more substantial?

Does the claim allow uncertainty and exceptions?

Am I persuaded because the evidence is good, or because the description feels uncannily like me?

And perhaps most importantly:

What am I planning to do with this information?

The amount of verification I need should change with the consequence.

A Reddit thread might give me a useful question to think about. A reel might give me a term to look up. An AI conversation might help me organise what I want to discuss in therapy. A podcast might introduce a theory that changes how I think about relationships.

None of those requires me to hand the source authority over my life.

If I am going to stop medication, conclude that I have a disorder, decide someone else has one, abandon treatment or make another consequential decision, the evidentiary bar needs to become considerably higher.

There is an astonishing amount of useful psychological knowledge online, mixed with oversimplification, advertising, personal testimony, serious scholarship, ideology, misunderstanding and nonsense. Unfortunately, they do not arrive neatly labelled.

So perhaps digital mental-health literacy isn’t learning which sources are good and which are bad. It is learning what kind of question each source can help answer—and noticing when we have asked it for more than it can know.