Dear Decision Maker

I need to talk about something I have been sitting with for a while.

And I am going to be honest with you. This is not comfortable to write.

The most consequential intelligence experiment of the 21st century is unfolding right now. And most people watching it are drawing the wrong conclusions.

The Machine That Never Runs Out of Targets

Here is what is happening in Gaza.

The IDF deployed three AI systems after October 7.

1. Lavender
2. Gospel
3. Where's Daddy

Lavender

Lavender ingests mass surveillance data on Gaza's 2.3 million residents.

• Cell phone metadata
• WhatsApp group memberships
• Phone change frequency
• Address history
• Social media activity

It assigns every individual a score from 1 to 100 representing the probability they are an active militant.

It marked approximately 37,000 Palestinians as suspected militants.

The overwhelming majority were assessed as junior operatives.

Gospel

Gospel handles infrastructure. Where human analysts previously produced roughly 50 targets per year in Gaza, Gospel generates 100 targets per day.

Where's Daddy

Where's Daddy tracks Lavender-flagged individuals and alerts operators when they enter their family homes.

Strikes were timed for when the target was at home with their family at night.

Here is the part that should make every intelligence professional stop.

Intelligence officers testified they spent approximately 20 seconds per target. One stated plainly that he had

Twenty seconds...

I have spent longer deciding what to have for lunch.

What the Numbers Show

A leaked classified IDF intelligence database, published in August 2025, listed approximately 47,653 names of Palestinians considered active in Hamas and PIJ military wings.

The IDF believed it had killed approximately 8,900 of them. At the time, the total dead in Gaza stood at roughly 53,000.

That yields 83% civilian deaths.

I want you to sit with that number.

For comparison.

The civilian death rate in Bosnia was 57%. Syria was 29 to 34%. The US war in Afghanistan was 8 to 12%.

83%...

A peer-reviewed study in Frontiers in Public Health estimated combatant deaths at 12.7% of the total.

The civilian-to-combatant killing index was 7.01.

In 2008-09 it was 0.61.

In 2014 it was 0.96.

The researchers concluded the 2023 conflict was:

And here is the part nobody wants to say out loud.

These are not failures of the system. This is the system working as designed.

When a machine generates 37,000 targets and operators spend 20 seconds confirming each one, the targeting process is not intelligence.

It is industrial production.

October 7 Proved the Technology Trap Is Real

Let me take you back to the beginning.

Israel completed a $1.1 billion high-tech barrier along its Gaza border in December 2021.

• Cameras
• Seismic sensors
• Radar
• Remote-controlled guns with AI-assisted aiming.

It was the most technologically advanced border defence system on Earth.

Hamas defeated it in minutes.

• Cheap drones destroyed the communication towers
• Snipers took out the remaining cameras
• Tractors bulldozed through the fence
• Paragliders flew over it

But the deeper failure was human.

Jericho Wall

Israel allegedly possessed Hamas's actual attack plan.

A 40-page document codenamed "Jericho Wall."

It described everything that subsequently happened:

1. Rocket barrage at the outset
2. Drones to knock out surveillance
3. Gunmen entering by paraglides and motorcycles
4. Plans to take 200 to 250 hostages.

Israeli intelligence first obtained versions of this plan around 2018. In May 2022, the Intelligence Directorate compiled a full report. It was briefed to the Chief of Military Intelligence and Southern Command.

Senior officials assessed it as Hamas's aspirational wish list. Not current capability.

They were wrong.

Female spotter soldiers at the Nahal Oz base had been reporting for at least three months that Hamas operatives were training daily near the fence.

• Digging holes
• Placing explosives
• Practising breach techniques

Spotters

One spotter was told she knew nothing.

Another was threatened with a court martial for talking "nonsense."

15 of those spotters were killed on October 7.

7 were taken hostage.

Unit 8200 had reduced round-the-clock coverage of Gaza. It stopped monitoring Hamas's handheld radios in 2022.

The specialised OSINT unit tracking Palestinian social media was disbanded in 2021.

On the Saturday morning of October 7, Unit 8200 was not operational near the Gaza border.

The Shin Bet's own investigation, published in March 2025, delivered the most damning admission.

Since Israel's 2005 disengagement from Gaza, the Shin Bet had failed to maintain a single intelligence source within Hamas's leadership.

Read that again.

The most surveilled piece of territory on earth. Billions in technology. And not one human source inside the decision-making circle of the organisation that attacked them.

AI Tells You What. HUMINT Tells You Why

AI is extraordinary at pattern recognition. It can process metadata on millions of people and find correlations that no human team could identify.

That is real.

That is valuable.

I am not here to tell you AI has no place in intelligence.

But here is what AI cannot do.

• It cannot tell you why someone joined that WhatsApp group.
• It cannot tell you whether the person who changed phones 3 times did it because they are an operative or because they are a teenager in a war zone.
• It cannot distinguish a Hamas fighter from his brother using the same device.
• It cannot assess whether a junior operative left the organisation six months ago.
• It cannot understand family obligation, tribal loyalty, religious conviction, or the deeply human desire to protect the people you love.
• It cannot tell you intent.

And in the targeting cycle, intent is everything.

A satellite can photograph a troop movement. A signals intercept can tell you who called whom. A metadata algorithm can tell you who looks like a militant based on their digital footprint.

But only a human source can tell you why those troops are moving and what comes next.

Only a human source can tell you whether an operation is aspirational or imminent.

Only a human source can tell you that the person Lavender scored at 85 actually left Hamas two years ago.

October 7 proved this. The technology worked perfectly. The sensors detected the breach. The cameras saw the paragliders. The signals were there.

But the system had no human sources to tell it that the exercise drills had become rehearsals for a real attack.

That the intentions behind the observable patterns had changed.

And then, in response to the greatest HUMINT failure in Israeli intelligence history, the system doubled down on the exact approach that failed.

What This Means for You

I want to be clear about why I am writing this.

This is not an abstract policy debate. The AI-versus-HUMINT tension is playing out right now in every domain you operate in. Whether you work in:

1. Government intelligence
2. Corporate security
3. Risk consulting
4. Journalism
5. Law enforcement
6. Military intelligence
7. Private investigations.

Every one of these fields is being told that AI will transform collection, analysis, and decision-making. And it will.

The question is whether the professionals in those fields understand what AI can and cannot replace.

Because here is the pattern I see.

Organisations fall in love with technology.

They reduce investment in human relationships, human networks, and human judgment.

They mistake data volume for intelligence quality. And then, when the situation changes, when the adversary adapts, when the thing that matters is not what happened but why it happened and what comes next, they discover they have no human sources to answer the question.

That gap is where intelligence failures live.

And that gap is where the professionals who understand both the technology and the human dimension become irreplaceable.

5 Lessons from Gaza That Apply to Every Intelligence Professional

1. AI processes data. HUMINT processes people.

An algorithm can tell you what is happening across a population. A human source can tell you what one decision-maker is thinking. In most operational contexts, the decision-maker's thinking is what you actually need.

Takeaway for you: If you are building collection capabilities and they are entirely technology-driven, you have a structural blind spot. It will not be visible until it matters.

2. Speed is not accuracy.

Gospel generates 100 targets per day. Human analysts produced 50 per year. The volume increase is staggering. The accuracy decrease is equally staggering. When your system optimises for speed, something else gets optimised away. In Gaza, that something was civilian lives.

Takeaway for you: In your own work, when a tool gives you more output faster, ask what it is giving you less of. The answer is almost always judgment.

3. The human in the loop is only as good as the time they are given.

Twenty seconds per target is not human oversight. It is a legal fiction. If your organisation uses AI to generate recommendations and then gives humans insufficient time to meaningfully evaluate them, you do not have a human-in-the-loop system. You have a machine with a signature block.

Takeaway for you: When you evaluate AI-assisted processes in your own work, measure not whether a human is in the loop, but whether they have the time, training, and authority to overrule the machine. If they do not, the loop is the machine's.

4. Your adversary studies your technology.

Hamas did not try to outcompete Israel's technology. They studied it, understood its dependencies, and built a plan to defeat it with tractors and cheap drones. The most sophisticated surveillance system on earth was neutralised by people who understood how it worked and what it could not see.

Takeaway for you: Any collection method that becomes routine becomes predictable. And anything predictable can be defeated. This is why diversifying your collection, and maintaining human sources that operate outside the digital environment your adversary can study, is not a luxury. It is operational security.

5. Cultural fluency is not optional. It is operational.

The reason Israel had no human sources inside Hamas's leadership is not because it lacked resources. It is because maintaining human relationships in environments shaped by family obligation, tribal loyalty, religious identity, and decades of conflict requires a depth of cultural understanding that technology cannot replicate.

Takeaway for you: The most powerful collection tool in any environment is a professional who understands the human terrain. Not as a module in a training course. As a lived operating reality.

Here Is What I Would Suggest as Your Next Step

If this newsletter made you think about the gap between what technology can see and what human intelligence can reveal, I want to hear from you.

Not a pitch. A conversation.

I want to understand where you are in your career, what you are trying to build, and whether what we are developing at Grey Dynamics Intelligence School can help you become the kind of professional who understands both sides of this equation.

Reply to this Email with "TARGETING"

I will personally review your background and tell you, honestly, whether our HUMINT Fundamentals programme is the right fit for you, or whether there is a better path.

No pressure. No scripts.

Just one intelligence professional talking to another.

Because that is how HUMINT actually works.

Ahmed Hassan
CEO-Founder Grey Dynamics
​Where headlines end, ground truth begins

PS: If you are looking for a HUMINT solution for your team or organisation please respond with "Ground Truth"

Featured Article Read More

Special Forces Read More

Leadership Analysis Read More

Intel Reports Read More

Read More

Read More

Want to Think Like an Intelligence Officer?

Most people scroll. Professionals structure.
The Intelligence Cycle Fundamentals Program teaches you how to analyse threats, map influence, and predict the next move—with zero prior intel background.

→ Join the program now and stop drowning in information. Start thinking like the few who make sense of it.​