Social Media and Intelligence: The Crisis of Authority

Concert crowd holding up smartphones to record, each phone a new source of information

Some of my comments from last week’s IntelCollab meeting in Washington DC.

Originally posted October 2009. Expanded October 2026 with the full argument from the talk, “Social Media and Intelligence: The Crisis of Authority,” and a look at how it played out.

The threat was never employees on Facebook

In 2009, most organizations thought the social media problem was a productivity problem. Staff were wasting time on Facebook. Someone might leak something on a blog. Security officers wrote policies, IT departments blocked sites, and everyone felt they had handled it.

My argument to the intelligence professionals at IntelCollab was that this missed the point entirely. The real threat social media posed to intelligence was not distraction. It was the de-legitimization of authoritative information.

That matters because of what intelligence is for. An intelligence function exists to move leaders to act: to change a budget, kill a project, enter a market, or brace for a threat. It can only do that if its information carries authority. A warning that sounds like one more opinion among thousands does not move anyone. So the crisis of authority is not someone else’s problem. It belongs to the intelligence profession itself.

What social media does well

None of this was an argument against social media. It does two things extraordinarily well.

  • Collaboration. People who would never have found each other can now share observations, test ideas, and build on each other’s work in real time.
  • A massive proliferation of sources. Ground truth that once took weeks to surface through official channels now appears in minutes, from people standing next to the event.
Close-up of a finger scrolling a social media feed on a smartphone

Both are real gains for anyone who collects information for a living. The problem is the side effect: every source starts to look equally valuable. A post from an eyewitness, a post from a paid propagandist, and a post from a bored teenager arrive in the same feed, in the same font, with the same share button.

For an analyst, that flattening attacks the step that comes before any judgment: source evaluation. Intelligence tradecraft has long graded a source’s reliability separately from the credibility of what it reports. Social media strips away the cues that grading depends on, such as who the source is, what access they had, and what their track record looks like.

Toffler saw the power shift coming

Alvin Toffler described the underlying shift two decades earlier. In Powershift (1990), he argued that the real revolution in information technology was not faster computers but the overturning of bureaucracy: information moving out of the hands of the people at the top who hoarded it and spreading across the whole organization.

Bundled stacks of freshly printed newspapers on a pallet, a symbol of traditional information gatekeepers

Social media finished that job. What it destroys is the gatekeeper’s monopoly: the editor, the agency, the expert, or the analyst whose position alone once guaranteed an audience. Much of that destruction is healthy. Gatekeepers were often wrong, slow, or self-serving.

But we still need authority

Here is the uncomfortable part. Big institutions are not always wrong, and culturally we still need authority. A decision-maker cannot read ten thousand posts before a board meeting. A crisis team cannot wait for consensus to form on a comment thread. Someone has to produce a judgment that is reliable enough to act on, and the decision-maker has to know whom to believe.

Analyst with head in hands facing three monitors of charts and a world map at night

So the question I put to the room in 2009 was not how to stop social media. It was how to maintain informational authority in the Web 2.0 era. I saw two answers.

Answer one: algorithmic authority

The first answer is to let a system assign authority. Google’s PageRank was the model: it judged a page’s importance by the structure of links pointing to it, in effect counting the web’s votes (Brin & Page, 1998). Clay Shirky wrote the best-known short treatment of the idea in November 2009, describing algorithmic authority as trusting the output of an unsupervised process that draws on many sources of uncertain reliability, with no person vouching for the result.

Algorithmic authority scales, and that is its great strength. Its weakness took longer to see: algorithms measure attention, and attention is not truth. A landmark study of the question, covering roughly 126,000 rumor cascades on Twitter from 2006 to 2017, found that falsehood spread farther, faster, deeper, and more broadly than truth in every category. True stories took about six times as long to reach 1,500 people. The researchers also found that people, not bots, drove the difference, because false stories were more novel and more emotionally charged (Vosoughi, Roy & Aral, 2018).

An authority system built on engagement will, left alone, reward exactly the information an intelligence analyst should trust least.

Answer two: personal relationships

The second answer is older than any algorithm: maintain personal relationships. A leader trusts an analyst whose record they know, who has been right when it mattered, and who has told them uncomfortable things before. That authority is earned one judgment at a time, and it cannot be faked by a share count.

Senior executive in conversation with a trusted advisor in a dimly lit office

This is slower and does not scale. It is also the form of authority that has survived every change in information technology so far.

The crisis of authority in 2026

Seventeen years later, the crisis I described has become the normal operating environment.

  • Social media became an intelligence discipline. In 2012, David Omand, Jamie Bartlett, and Carl Miller made the case for social media intelligence, or SOCMINT, as a distinct discipline, and argued that it needed a public legal basis and clear rules to be legitimate.
  • Open source moved to the center. The U.S. intelligence community now treats open-source intelligence as a core discipline, set out in the IC OSINT Strategy 2024–2026 released by ODNI and CIA.
  • Synthetic content made the flattening worse. Generative AI can now produce plausible text, images, and accounts at almost no cost. The proliferation of sources that looked like pure gain in 2009 now includes sources that do not exist.

Abundance also feeds the analyst’s own weaknesses. When every position has a thousand supporting posts, confirmation bias never runs out of material. (I cover the mechanics in Cognitive Bias: A Field Guide.)

How an intelligence function keeps its authority

The answer I gave in 2009 still holds: intelligence professionals have to change the culture. In practice, that means a few disciplines.

  1. Grade the source before the content. Ask who is speaking, what access they had, and what their record is before deciding whether the claim is interesting.
  2. Separate collection from judgment. Social media is a superb collection environment and a terrible place to form conclusions. Gather there; decide elsewhere.
  3. Say what would change your mind. Judgments that state their own falsification conditions earn trust, because the decision-maker can watch them being tested.
  4. Keep score. Authority that rests on a visible track record is harder to drown out than authority that rests on a title.
  5. Invest in the relationship. The leader who knows and trusts their analyst will hear a warning over the noise. The one who has never met them will not.

None of this is quick or easy. But as I told the room at IntelCollab, it will be interesting.

Frequently asked questions

What is social media intelligence (SOCMINT)?

Social media intelligence is the collection and analysis of information from social media platforms to support decisions. David Omand, Jamie Bartlett and Carl Miller made the case for it as a distinct intelligence discipline in 2012.

What is algorithmic authority?

Algorithmic authority is trust placed in the output of an automated process, such as a search ranking or a recommendation feed, rather than in a named person or institution. Its strength is scale; its weakness is that it tends to measure attention rather than accuracy.

How should analysts judge the credibility of social media sources?

Grade the source before the content: who is speaking, what access they plausibly had, and what their track record is. Then corroborate the claim independently before treating it as evidence.

Want this kind of analysis for your organization? Contact Competitive Futures or learn more about Eric Garland.

About Competitive Futures

Competitive Futures has provided strategic intelligence and foresight to corporate and government clients since 2002. Its principal, Eric Garland, publishes analysis here, on Game Theory Today, and in the Systemics Quarterly Brief.

Read next

Stay in touch

Leave your email and a note about the Brief, a briefing for your team, or this piece, and we will be in touch.

Similar Posts