DEV Community: Moustafa hamwi

LinkGuard: How to Tell If a Link Is Suspicious Before You Click

Moustafa hamwi — Sat, 24 Oct 2026 21:02:21 +0000

Originally published at scouthub.dev.
LinkGuard is a tool for checking suspicious links before you open them. You paste the link, and the tool follows the chain of redirects behind it, asks five global threat-intelligence sources at the same time, then gives you a risk score from 0 to 100 with a plain explanation of why. It is open source under the MIT licence, and its website works in Arabic and English.

Its motto is simple: know before you click.

Why do we need to check links?

Phishing is a crime that combines social engineering with technical tricks to steal people's personal data and financial account details, according to the definition used by the Anti-Phishing Working Group (APWG) [1].

It is not going away. In 2025 the APWG observed about 3.8 million phishing attacks, up slightly from 3.76 million in 2024 [1]. Its report for the last quarter noted that the use of text messages for phishing rose [1].

The link is a common doorway for these attacks: a page that looks like your bank or your account, but whose address differs by one letter.

How does LinkGuard check a link?

It follows the redirects. Many suspicious links pass through several redirects before they reach their real destination. The tool follows them for up to 8 hops. It also protects itself against SSRF attacks, which abuse an application to reach the internal network or the machine itself [3]. That protection is necessary for any tool that opens links sent by strangers.
It asks five sources at once. They are VirusTotal, Google Safe Browsing, URLhaus, PhishTank and AbuseIPDB. VirusTotal itself brings together the verdicts of more than 70 antivirus scanners and blocklist services [2].
It runs its own checks. It detects look-alike domains and similar-looking characters, and checks the domain's age and SSL certificate, with no outside API keys needed.
It calculates the score and explains it. From 0 to 100, with a "Why this verdict?" panel beside it that lists the reasons.

What does the score mean?

The score is not a mystery number. Every verdict comes with the reason for it, so you understand it instead of just trusting it.

If one of the five sources fails during a scan, confidence in the result drops and the scan keeps going. In other words, the tool gives you the best it has and is honest that it did not see everything.

No tool guarantees 100%. The website itself says results may not be fully accurate, and that scanning a link does not mean endorsing it. Use the tool as an extra layer of protection, not as a replacement for caution.

What else is in the tool?

A QR code scanner, so you can check a code before you open what is behind it.
A scan history stored on your own device.
An installable web app (PWA) for your phone, with an Android app coming soon.

Five habits that protect you before you click

Look at the real domain, the part before the first single slash in the address, not the name shown in the message.
Be wary of urgency. "Your account will be closed within an hour" is a phrase scammers use to stop you from thinking.
Don't enter your details from a link that reached you in a message. Open the official site or app yourself.
Be careful with shortened links and QR codes from a source you don't know.
Check the link with a tool before opening it, and if doubt remains, don't click.

Why did we open the code?

A tool that judges people's links should let people see how it judges. That is why it is licensed under MIT. It is one of ScoutHub Lab's initiatives.

Frequently asked questions

What exactly does LinkGuard check?

It follows a link's redirects, asks five sources at once, checks for look-alike domains, the domain's age and its certificate, then gives you a score from 0 to 100 with an explanation.

Can I rely on it alone?

No. No tool guarantees full accuracy. Combine the tool with the habits of caution listed above.

What happens if one of the sources fails?

Confidence in the result drops and the scan continues.

Does it scan QR codes?

Yes, it has a QR code scanner.

Does it work on a phone?

Yes, as an installable web app. An Android app is coming soon.

Is the tool open source?

Yes, under the MIT licence.

Where can I find it?

At linkguard.scouthub.dev, and there is a short profile on the project page.

Smart Scout Manager: How to Run Your Scout Troop or Association from One Place

Moustafa hamwi — Sat, 24 Oct 2026 21:01:59 +0000

Originally published at scouthub.dev.

Smart Scout Manager is a platform that brings the running of a scout troop into one place: members, activities, finances, equipment and reports. AI drafts the paperwork, such as minutes, reports and plans, and the leader reviews it and decides. There is a free version, and a separate version for associations and councils.

We built it at ScoutHub Lab for a simple reason: we saw leaders spending their time on paperwork instead of with the kids.

Why do scout troops struggle with paperwork?

Scouting is a large movement. The World Organization of the Scout Movement says it has more than 60 million active members in 176 countries and territories [1].

Behind every troop is a volunteer leader. After the meeting and the activity, more work waits: member lists, attendance, the cash box, equipment, and the report at the end of the term and the year.

When a new leader takes over, many details get lost because the records lived in a notebook or on one person's phone. So we chose to keep everything in one record that stays with the troop even when leaders change.

What does the platform manage?

Members and leaders. Digital records and printable ID cards that can be issued in one click.
Programme. A dashboard that shows the troop's activities and statistics on one screen, with annual programme planning and goal indicators.
Money and equipment. The cash box, expenses and equipment, with records you can audit.
Minutes and reports. Ready-made admin forms, quarterly and annual reports, meeting minutes and instant certificates.
Scouting knowledge. A skill-badge reference, a library of traditions and ceremonies, and child-protection guides.
Parents and the troop's public face. A parent portal, and a public troop website that shows achievements and welcomes new members.

Some of these features belong to the paid plans. The details are on the platform's page.

How does AI help without replacing the leader?

AI in the platform writes drafts. It helps with more than 12 tasks, including minutes, reports, plans and design material. The leader reads, edits, then approves. The decision is theirs, not the tool's.

We hold to this rule in everything we build: the tool gives the leader time back, and the leader spends it with the kids. We wrote about this idea in detail in Human First.

What does it offer associations and councils?

Associations and councils have a separate Enterprise plan, priced to fit the need. It adds dashboards across several troops, consolidated reports, badge approval, a training-bag generator, a scout nickname system, circulars, financial projects and media management.

The idea is that an association's officer needs one picture of what is happening across all troops, instead of collecting it from scattered reports.

How do we protect members' data?

Troop members are mostly children and teenagers, so we handle their data carefully. The scouting library inside the platform also includes child-protection guides.

Google Calendar sync is one example of what we do:

It is optional, and it asks for the narrowest permission possible: access only to the calendar the app itself creates.
The app cannot see your other calendars, edit them or delete their events.
The access token stays in your browser and is never sent to our servers.
We do not store calendar data, share it, sell it, or use it for advertising or for training AI.
You can revoke access from your Google account at any time.

We follow Google's API Services User Data Policy, including its "Limited Use" requirements, which prohibit transferring user data to advertising platforms [2]. In the database we use PostgreSQL Row Level Security.

How do you start?

Go to ssm.scouthub.dev. The platform works in Arabic and English, on computers, phones and tablets (Windows, Mac, Android and iOS).

The free version includes member and leader records, admin forms, the badge reference, equipment and cash logs, and the full scouting library. From there you can move to the plan that fits your troop or association. You will also find a short profile on the project page.

Frequently asked questions

Is Smart Scout Manager free?

There is a free version with the core features listed above. The paid plans add things like the parent portal, AI planning and unlimited records, and associations have their own plan.

Does it suit large associations and councils?

Yes. That is the purpose of the Enterprise plan: dashboards across troops, consolidated reports, circulars and badge approval.

Does AI replace the leader?

No. It writes drafts, and the leader reviews them and decides.

Does the platform read all of my Google Calendar?

No. It asks for access to one calendar that it creates itself, and it cannot see your other calendars.

Which languages does it support?

Arabic and English.

Who builds the platform?

ScoutHub Lab, a digital innovation lab that builds platforms for scouting, education and community work.

Najem: The AI Scout Assistant on Telegram for Planning Activities and Camps

Moustafa hamwi — Sat, 24 Oct 2026 20:56:25 +0000

Originally published at scouthub.dev.

Najem is an AI scout assistant that works inside Telegram. You describe your activity or camp, and it gives you a detailed plan with a schedule, an equipment list, safety guidelines and an evaluation method. It is free, needs no sign-up or credit card, and is built on the Muslim Scouts and Lebanese scouting curricula.

Najem suggests. The leader decides and leads.

What is Najem, and who is it for?

Najem is for scout leaders in the Arab world, especially those working with the Muslim Scouts or Lebanese Scouting curriculum. It recognises the terminology of each curriculum and adapts to it. It also has dedicated assistants for the troop administrator, the commissioner's advisor and the leader trainer.

The World Organization of the Scout Movement says scouting has more than 60 million active members [1]. According to the platform's own figures, Najem has generated more than 5,500 activity plans and served more than 1,350 leaders in 12 Arab countries.

How do you start?

Choose the curriculum: Muslim Scouts or Lebanese Scouting.
Open the bot on Telegram. Telegram describes bots as small applications that run entirely inside its app [2]. If you already have Telegram, there is no new app to install.
Describe your activity directly, or use the prompt generator on the website.

No registration and no credit card are needed.

What does Najem offer?

Activity and camp plans with a detailed schedule.
Activity suggestions that suit each age stage.
Equipment lists, safety guidelines and an evaluation method.
Turning the scout values, promise and law into practical behaviours.
Quick educational ideas for changing negative behaviour and building teamwork.
Support for skills such as camping, navigation, leadership, communication and problem-solving.
Help with meeting minutes, digital records and periodic reports.
For commissioners: planning strategic goals and performance indicators.

An assistant for every stage and role

Cubs and Girl Cubs: stories, games and basic skills.
Scouts and Guides: the patrol system, leadership, and proficiency and progress badges.
Rovers: community service projects and exploration planning.
Troop administrator: an administrative assistant for daily work.
Commissioner's advisor: planning at council level.
Leader training: including preparation for the Wood Badge course.

The prompt generator: for when you don't know where to start

If you are not sure how to word your request, open the prompt generator on the website. It asks for the goal, the age stage, the group size and the place, then prepares a text that you send to Najem.

It has ready templates for a camp, a community service project, a leadership course, a day trip, a cultural activity and a campfire night.

Consult it together with your group

Najem can be added to the leaders' Telegram group as an admin, so the whole group can ask it questions and generate activities together.

We prefer that leaders discuss a plan among themselves before they adopt it, because sitting together around the plan is part of building a team. We wrote about this idea in Human First.

Najem suggests, the leader decides

A plan generated by Najem does not replace the leader's experience, supervision or the field safety rules. This is written on its website, and we repeat it here.

Review every plan, adjust it to your kids, your place and the weather, and apply the safety rules your troop follows.

Frequently asked questions

Is Najem free?

Yes. It is free and needs no sign-up or credit card.

Which curricula does it support?

Muslim Scouts and Lebanese Scouting, and it recognises the terminology of each.

Can I use it in a Telegram group?

Yes. It is added to the leaders' group as an admin.

Does it write minutes and reports?

It helps with meeting minutes, digital records and periodic reports. For full troop management, see Smart Scout Manager.

Is it useful for commissioners?

It includes a commissioner's advisor that helps plan strategic goals and performance indicators.

Where is the project page?

Here, with the link to Najem's website.

Teaching AI to Today's Generation: Why, From What Age, and How to Start at Home, in School and in the Scout Troop

Moustafa hamwi — Sat, 24 Oct 2026 20:56:04 +0000

Originally published at scouthub.dev.

Yes, children and teenagers should learn about AI. Not to become programmers, but to understand what they use: how it works, when it gets things wrong, and what happens to their data. The best places for this learning are the home, the school and the scout troop, with adults who sit beside them rather than only banning the tools or leaving kids alone with them.

In short:

Children are already using these tools, and their parents mostly don't know.
"Teaching AI" here means understanding, checking and protecting data, not memorising tool names.
Unguided use can harm learning, and guided use reduces that harm.
The first step is a short session with the child, not a big decision.

Why should children learn about AI?

Because they already use it, whether we teach them or not.

In a Pew Research Center survey of 1,391 U.S. teens aged 13 to 17 in autumn 2024, 26% said they had used ChatGPT for schoolwork. A year earlier the figure was 13%. It had doubled in a single year [1].

And this isn't one school's story. A Common Sense Media report from September 2024 says seven in ten young people have used at least one generative AI tool, most often for schoolwork. More important than the number is this line: most parents aren't aware that their teens are using it at all [2].

So the picture is clear. Children got to the tool before we did. If they don't learn to use it with us, they will learn alone, and nobody will ask them: is this answer correct?

The adults are not ready yet either. Stanford's 2025 AI Index reports that 81% of computer science teachers agree that using and learning about AI should be part of foundational computer science education. But fewer than half of high school CS teachers said they felt equipped to teach it [3].

There is a gap between what teachers want and what they feel ready for. A teacher alone can't close it. It takes parents, leaders and youth organisations moving alongside.

What do we mean by "teaching AI"?

We don't mean memorising tool names or learning magic phrases for writing prompts. Those change every few months.

We mean something closer to literacy for a new era. A child understands the basic idea, judges the result, and knows their own limits and the tool's limits.

The world has begun to give this a clear shape. In 2024 UNESCO published its AI competency framework for students. It sets out 12 competencies across four dimensions: a human-centred mindset, the ethics of AI, AI techniques and applications, and AI system design. Each competency has three levels: understand, apply, create [4].

UNESCO says that integrating AI learning objectives into official school curricula is crucial, so that students engage with the technology safely and meaningfully [4].

A second framework is called AI Literacy (AILit). It is a joint initiative of the OECD and the European Commission for primary and secondary education, and its final version was launched on 18 June 2026. It divides learning into four domains: engaging with AI, creating with AI, managing AI, and shaping AI. It gathered feedback from more than 2,000 people in more than 100 countries, and it contributes to the innovative domain of the 2029 PISA assessment [5].

For adults, Article 4 of the EU AI Act requires providers and deployers of AI systems to take measures to ensure, to the best of their ability, a sufficient level of AI literacy among the people working with them. It has applied since 2 February 2025 [6]. The message is that AI literacy is now treated as a basic skill for adults too, not a luxury.

We boil all of this down to four questions any child can learn to ask: What is this system doing? Where did its answer come from? Should I believe it? And who is responsible if it is wrong?

Why is this an upbringing issue before it is a tech issue?

Because the hard questions here are questions of values. Do we say so when we hand in work we did with a tool's help? Do we check before we share? Do we type a friend's details into a program when we don't know where the data goes?

Upbringing has known these questions long before AI appeared: honesty, trustworthiness, responsibility for what we say and do. It is not strange to UNESCO's framework either, since the first two dimensions in it are a human-centred mindset and the ethics of AI [4]. And we saw that teens distinguish between a use that teaches them and a use that is cheating [1].

So upbringing is not late to this topic. It already has something to say. Scouting, for one, has taught honesty and trustworthiness for more than a century. What we lack is the link between those values and the screen in a child's hand today.

Does AI harm children's learning?

It can, when used without guardrails. A field experiment published in PNAS in 2025 showed this.

The experiment covered nearly a thousand high school students in Turkey studying mathematics. Some got a tool that works like ordinary ChatGPT. Others got a "tutor" built with safeguards, which gave hints instead of handing over the solution [7].

The result deserves a pause. During practice with the tool, the ordinary-tool group's scores rose by 48%, and the guarded-tutor group's by 127%. But in a later exam with no tool at all, those who relied on the ordinary tool scored 17% lower than students who had never used it. The guarded-tutor group scored close to the group that used nothing, which means the safeguards largely reduced the harm [7].

The lesson is not "ban the tool". The lesson is that a tool that hands over the answer gives the child temporary speed and takes away the practice. A tool that asks and hints keeps the practice. The researchers warned that unguided use can become a "crutch" that prevents the basic skills from forming [7]. It is one experiment in one country, so we read it with care and don't generalise. But it says something worth taking seriously.

And the risk isn't limited to learning. There are three more that both children and adults should know:

Confident mistakes. A tool can give a wrong answer in a confident tone. So we don't learn from it without checking another source.
Privacy. UNICEF's guidance on AI and children, in its third version from December 2025, sets out ten requirements for child-centred AI, grounded in the Convention on the Rights of the Child. They include protecting children's data and privacy, their safety, and preparing and equipping children to deal with AI [8]. This is why we never type a child's name, school or photos into a tool.
Cheating. Teenagers themselves are not of one mind. In the Pew survey, 54% said using ChatGPT to research a new topic is acceptable. Only 18% said using it to write an essay is acceptable, while 42% said it is not [1]. Young people tell apart a use that teaches them from a use that does the work for them. We should build on that instinct, not confiscate it.

From what age do we start?

No single age is agreed by the sources, and we are wary of anyone who gives you a firm number. What we can say is that the modern frameworks were designed for primary and secondary school together, and that they progress from understanding to applying to creating [4][5].

What follows is our practical reading. It is our suggestion, not a research finding:

Young children: they don't need a chatbot. Simple games that explain the idea are enough, such as sorting things into groups and understanding that machines learn from examples.
Middle years: a short try-out next to an adult, with a question after every answer: how do we check it?
Teenagers: more independent use, with a clear agreement on what is allowed for schoolwork and what isn't. And read each tool's terms, since each has its own policy on permitted age.

What can parents, teachers and scout leaders do this week?

We suggest small steps that can start tomorrow:

Sit with them and try it. Ask the tool for something simple in front of the child, then ask: do you believe this answer? Why?
Test one mistake. Ask the tool about something you know well, such as your neighbourhood's history or a rule in a scout game. Watch together to see whether it slips.
Try first. Agree that the child attempts the problem alone before asking the tool. If they do use it, they ask for a "hint", not "the solution". This is the difference the PNAS experiment showed [7].
Write a short rule. At home, in class or in the troop: what do we use the tool for? What do we write by hand? When do we say that we used it?
Protect data. No full names, no addresses, no photos, and no information about the troop or the families [8].
Make it a group activity. In the scout troop, hold a session called "Spot the Mistake". Each scout brings an answer from a tool, and everyone looks for its error or for proof that it is right. It brings them together, teaches checking, and needs no equipment.

A teacher or leader doesn't need to be an expert to start. The AILit framework is aimed at teachers, parents, education leaders and curriculum designers together [5]. Being curious is enough, along with being willing to tell a child: I don't know, let's check together.

Questions parents often ask

Does AI make children less intelligent?

The sources we read don't say that. What the PNAS experiment says is narrower and more precise: students who relied on a tool that hands over ready answers did worse on an exam without the tool, while those who used a tool with safeguards did not [7]. The problem is how it is used, not the child.

Should we ban it at home?

A ban alone may push the child to experiment out of our sight. We also saw that most parents don't even know what their children do with these tools [2]. Better is an open agreement, an occasional shared session, and knowing what the child uses.

Does a child need to learn programming?

Not necessarily. The frameworks we mentioned start with understanding and engaging, and later reach creating and design [4][5]. Programming is valuable for those who love it, but it isn't a precondition for a child to be aware of what they use.

Is it enough to teach them how to write prompts?

No. Good prompt-writing changes with every new tool. The question "should I believe this answer?" stays valid with any tool.

Where we stand

At ScoutHub Lab we build platforms for education and scouting, and we design them so that the human stays the decision-maker.

What we wish for our children is not that they fear the tool, nor that they hand it their minds. We wish they grow up asking: how does it work? Should I believe it? And what is my part?

Code and a Code of Honor: Why Tomorrow's Technology Needs Firm Educational Values

Moustafa hamwi — Sat, 24 Oct 2026 20:50:29 +0000

Originally published at scouthub.dev.

Every scout camp has a quiet moment we love. The young people stand in a circle, and one of them raises a hand to say the Promise for the first time. It's only a few words, but they know those words commit them to something. They know someone will ask them about it tomorrow.

We write code during the day and sit in these circles in the evening. Over time, one question kept coming back to us: why do we ask a twelve-year-old to keep a code of honor, but ask nothing like it of the software they talk to every day?

The tool is no longer neutral

We used to say technology is just a tool, and what matters is how we use it. That's no longer enough. AI doesn't simply wait to be used. It suggests, answers, reassures, and sometimes decides.

And young people are already using it. In a Pew Research Center survey of 1,458 U.S. teens, 54% said they had used chatbots for help with schoolwork, and 12% said they had turned to them for emotional support or advice [6].

That second number deserves a pause. A teenager looks for advice in a vulnerable moment and finds a machine answering. What will it say? And who decided?

In a scout troop, we don't let a young person share their worries with someone we don't know. We choose leaders carefully, train them, and set clear rules for every meeting. Yet the machine that same young person talks to alone in their room at night has been vetted by none of us, and we don't know who wrote its instructions.

Another study, from Common Sense Media, found that 72% of teens had used an AI companion at least once. About a third of teen users said something an AI companion said or did made them uncomfortable [5].

In the wider picture, the AI Incident Database recorded 233 incidents in 2024, a 56.4% increase over the year before, according to the Stanford AI Index [4].

These numbers aren't a reason to fear technology. They mean the tool now carries values, whether we like it or not. The real question is: whose values?

The world is writing rules. What about us?

The good news is that the world has started writing rules. In 2021, UNESCO's 193 member states adopted the first global standard on the ethics of AI [1].

In December 2025, UNICEF published version 3.0 of its guidance on AI and children. It sets out ten requirements, including children's safety, their data and privacy, transparency, and their best interests and development [2].

The EU AI Act goes further. Its Article 5 prohibits AI systems that exploit people's vulnerabilities due to their age, and that ban has applied since 2 February 2025 [3].

These texts matter. But we know from our work in education that a text alone doesn't shape behavior. A law tells you what you may not do. Education builds in you the reasons you wouldn't want to do it in the first place.

This is where we think the Scout Movement holds something the tech industry needs. World Scouting engages more than 60 million young people and volunteers, across 176 National Scout Organizations [7]. All of them started at the same point: a short promise and a clear law.

What the Promise and Law teach a programmer

The Scout Promise doesn't open with a list of prohibitions. It opens with a commitment. And the Scout Law doesn't say "don't lie." It says a scout is trustworthy, and that their honor is to be trusted.

The difference is small in words and large in design. When we build a product on a list of prohibitions, we bolt on the limits at the end, like a patch. When we build it on a commitment, values become part of the very first decision.

Take trust. A scout is trusted because they don't take more than they need and don't claim what isn't theirs. That's exactly what we want from any tool that deals with young people.

When we built the Google Calendar integration in "Smart Scout Manager," we could have asked for full access to a leader's calendar. That would have been easier to code. Instead we chose the narrowest scope available, which only manages calendars the app itself creates (calendar.app.created). We don't see the leader's private appointments, and we don't need to.

On the same platform we included child-protection guides, because a leader running a group of children needs those before any other feature.

With "eNaawah" we made a similar call. User data stays in the browser, and we don't ask anyone to create an account. Someone writing an obituary notice for a loved one shouldn't have to hand us their email address.

Letting AI know its limits

There's a meaning in the Scout Law we keep coming back to: know your role and do it well. A leader doesn't do everything for the young people. They guide, and leave room for them to learn.

We apply this to AI itself. In "Kashef," our website security audit platform, the AI doesn't compute the security score. The score comes from fixed rules that can be reviewed and repeated, so the same site gets the same result every time.

The AI's job there is to explain. It takes the technical result and turns it into language the site owner understands. That's a useful role, but it isn't the referee's role.

Why does this matter to us? Because a language model can give you two different answers to the same question. That's acceptable in an explanation and unacceptable in a verdict. When we mix the two roles, we lose trust.

We do the same with "Najem," our AI assistant on Telegram. We built it on the curricula of the Muslim Scouts Association and Lebanese scouting, so it turns the Promise, the Law and scout values into practical activities a leader can run with their group. Today more than 1,350 leaders in over 12 countries use it.

But Najem doesn't raise children. The leader does. Najem might suggest an activity on honesty and explain how to run it, then leave it to the leader to read the faces in front of them and decide. The tool serves the educator; it doesn't replace them.

There's another reason we built it on the curricula rather than on the model's general knowledge. A language model learned from the whole internet, the good and the bad. When a leader asks it about a value like selflessness, we want the answer to come from an educational source reviewed by leaders and educators, not from the first text the model happened to see. That way, the values Najem offers stay the same values we teach at camp.

A small charter for anyone building tech for young people

We don't claim to have the full answer. But we've gathered what we've learned into a few principles, and we try to keep them the way we keep the Promise:

Ask for the least you need. Every extra permission is trust nobody gave you.
Know the AI's role. It explains, suggests and assists. Decisions that need consistency and fairness stay with clear rules or with a person.
Put child protection before features. If children use the platform, the safeguarding guide ships in version one, not version ten.
Keep the human educator at the center. The tool serves the leader, the parent and the teacher; it doesn't go around them to the child.
Be honest about what the tool is. Don't let it pretend to be a friend or a person.
Leave data with its owner whenever you can. What you don't collect can't leak.
Build values from the source. If you have trusted educational curricula, make them the knowledge base. Don't let the model improvise values.

None of these principles is new. Most of them appear in the UNICEF and UNESCO guidance in other words [1][2]. What's different is that we see them from the campsite: a daily commitment, not a document pinned to a wall.

From the circle to the repository

What we suggest is simple. Before you write the first line of a project aimed at young people, write the product's "promise." Three or four sentences stating what the tool commits to for the people who use it.

Put it in the first file of the repository. Return to it whenever you add a feature, request a new permission, or give the AI a new role. Ask yourself: does this honor the promise?

If you work with a team, read it together, the way a scout troop reads its Law. It will feel odd at first, then it becomes a habit, and then it becomes a culture.

Good code works. Code that carries a code of honor deserves to be trusted.

The Prompt Generation: How Do We Teach Young People to Ask Deep Questions Instead of Consuming Ready-Made Answers?

Moustafa hamwi — Sat, 24 Oct 2026 20:50:07 +0000

Originally published at scouthub.dev.

A fourteen-year-old sat down to do homework about a polluted river near his home. He typed into the chat window: "Write me an essay about river pollution." A few seconds later he had five tidy paragraphs. He handed them in the next day. His teacher asked: what surprised you most while writing this? He had no answer.

This scene is no longer rare. In the United States, 26% of teens aged 13 to 17 said they had used ChatGPT for schoolwork in 2024, double the 13% of 2023 [1]. In our own Arab homes and schools we see the same trend, even if we don't yet have precise numbers.

The problem isn't the tool. The problem is the question we put to it, and what we do once the answer arrives.

What happens to a mind that takes answers ready-made?

At the MIT Media Lab, 54 participants wrote essays over several sessions. One group used a language model, one used a search engine, and one wrote with their own brains only. The researchers measured brain activity with EEG. The weakest neural connectivity was in the language-model group, and many of them couldn't accurately quote what they had written themselves minutes earlier [2]. The researchers called this "cognitive debt." The sample is small and the findings are preliminary, but they look a lot like what we see in classrooms.

In a study by Microsoft and Carnegie Mellon University, 319 knowledge workers described 936 real situations from their work. The researchers found a clear pattern. The more confidence people had in the AI, the less critical thinking they did. The more confidence they had in themselves for the task, the more critical thinking they did [3].

This matters for education. A child who doesn't trust their own ability will hand their mind over to the machine easily. So building self-confidence isn't only an emotional matter. It's a condition for thinking.

And that study looked at adults in their jobs, not children in school. If an experienced professional relaxes in front of the tool, what about a child who has no experience yet to measure it against?

The same tool can teach or weaken

We don't want to say that AI is the enemy of learning. The evidence says something more precise.

In an experiment with nearly a thousand high school students in Turkey, researchers gave one group a standard version of GPT-4 and another group a version designed as a tutor that doesn't hand over the solution. During practice, the first group's grades improved by 48% and the second by 127%. But when the tool was taken away for the exam, the first group did 17% worse than students who had never used it. The "tutor" group was largely unharmed [4].

At Harvard, physics instructors designed an AI assistant that keeps its replies short, reveals one step at a time, and asks the student to try first. Students who learned with it gained more than twice as much as in the classroom session, in less time [5].

The difference between the two experiments isn't the power of the model. It's the design: does the tool give the answer, or hand the question back to its owner?

From rote learning to meaning engineering

We named one of our projects "Talqeen Dev." The name carries a deliberate irony. In our school memories, talqeen means having the answer dictated to you so you can memorize it. But the tool does the opposite. It takes a developer's raw idea in Arabic and turns it into a plan and small milestones. To work well, the developer has to be clear about what they want, for whom, and within what limits. In other words, it teaches them how to ask, not what to memorize.

That's what we mean by "meaning engineering." A good question isn't a trick for squeezing a better answer out of a machine. It's practice in organizing an idea before asking for anything. When you write a clear request, you are really analyzing the problem: what are the givens? What do I already know? What's missing? How will I know the answer is right?

Let's put two examples side by side.

The weak question: "Write me something about river pollution."

The deep question: "There's a small river in my area that turns dark in summer. What are three possible causes? How can I check each cause myself with simple tools? Don't write the essay for me. Help me organize my thoughts."

The second question doesn't need genius. It only needs the boy to have thought before typing. That's where real learning begins.

Another example, from the world of leaders. The weak question: "Give me a game for scouts." The deep question: "I have twelve boys aged 11 to 13, in a small schoolyard, and we want a game that trains communication inside the patrol. What could go wrong, and how will I know it met its goal?" The first question gets you an activity. The second makes you understand your activity.

So why is rote learning still so strong? Because our exam systems reward the right answer, not the path to it. Students are smart. They learn quickly what gets rewarded. If the exam only asks "what?", nobody will bother with "why?" and "how?". AI didn't create this problem, but it exposed it. A tool that remembers everything has made memorizing alone almost worthless.

What we've learned from scouting and from our tools

For more than a century, the scout method has rested on learning by doing. We don't explain to young scouts how to pitch a tent and then give them a written test. We hand them the tent and the ropes, let them make mistakes, and then ask: why did it fall? The patrol system makes a small group discuss, decide, and own the result. That's a natural environment for asking questions.

When we built the game "ScoutQuest" (درب الكشاف), we put in 1,056 questions across eight categories, with daily challenges and one-on-one duels. We should be honest: a quiz game can end up training recall only. "What color is the Cubs' scarf?" is a memory question. "Your patrol is lost in the woods with one hour of daylight left. What are your first three steps?" is a question that makes a child think and set priorities. The challenge we're still working on is adding more of the second kind.

In "Najem," our AI assistant for scout leaders, the prompt generator doesn't just ask "what do you want?" It asks the leader about the goal, the age stage, the number of participants, and the place. A leader who answers those four questions has already done half the planning before seeing any answer. We've noticed some leaders now ask themselves these questions even without the tool.

Global numbers say these skills aren't a luxury. In the PISA 2022 creative thinking assessment, in which 64 countries took part, 22% of students across OECD countries did not reach the baseline level [6]. In the World Economic Forum's Future of Jobs Report 2025, analytical thinking remained the top core skill for employers, at 69% [7]. And in 2024 UNESCO published an AI competency framework for students with 12 competencies across four dimensions and three levels [8]. Its common thread is that students should understand the tool and question it, not just use it.

Five exercises for parents, leaders and teachers

We don't need new curricula to start. We need small habits, repeated:

The question before the screen. Before opening any tool, the child writes on paper: what do I know about this topic? What do I want to know?
Improve the question three times. They write their question, then rewrite it with more context, then add a constraint. Then we compare the three answers.
Catch the mistake. Ask the tool for an answer, then challenge the child to find a missing or wrong fact in it and prove it from another source.
Explain it in your own words. After every answer from the machine, they explain it to a younger sibling or their patrol in two minutes. If they can't, they haven't understood it yet.
The best question wins. In the scout meeting or the classroom, reward the best question, not the fastest answer.

What we want from the prompt generation

We don't want a generation that is only good at writing commands for machines. We want a generation that understands that writing a good request is, at heart, a way of thinking: defining the goal, understanding the context, doubting the first answer, and going back to the question.

What's asked of us is simple: slow down a little before the answer, and ask our children "how did you get there?" more often than "what's the answer?"

Answers have become cheap. A good question still needs a person who thinks.

When Algorithms Think for Us: Who Shapes Humanity's Future, Us or Our Smart Models?

Moustafa hamwi — Sat, 24 Oct 2026 20:42:24 +0000

Originally published at scouthub.dev.

We write these lines as people who use AI every day. We build tools with it, review text with it, and test ideas with it. So we are not writing from fear, and not from awe. We are writing as people who see the tool up close and ask a simple question: as the tool gets smarter every year, who remains the one who decides?

This is not only a philosophical question. It is a practical one. The teacher grading homework written by a language model faces it. So does the doctor reading an automated recommendation, and the developer accepting a suggestion they didn't fully read. And we face it in the lab, too.

A map that changes every few months

Let's start with what we know, not what we imagine.

In a single year, model scores on hard benchmarks jumped sharply. Stanford's AI Index reports a gain of about 67 points on a benchmark of real-world software problems (SWE-bench) within a year of its release [2]. The same report notes that the cost of running a GPT-3.5-level system fell more than 280-fold between late 2022 and autumn 2024 [2]. What was expensive and rare became cheap and everywhere.

Adoption has kept pace. According to the same report, 78% of organizations said they used AI in 2024, up from 55% the year before [2]. These are no longer tools for big labs only. They sit in the office, the classroom and the phone.

And it's not just accuracy that is improving, but stamina. In a 2025 study, METR found that the length of tasks AI agents can complete with 50% success has roughly doubled every seven months over the previous six years [3]. METR itself notes that some of its figures age as new data comes in. But the direction is clear: the machine no longer just answers a question. It carries out a chain of steps.

Behind all of this is computing power growing fast. Epoch AI estimates that the training compute of notable models has doubled roughly every six months since 2010 [4].

What does this mean for us? It means today's map won't hold in a year. Anyone planning a school or an organization around what models can do right now will find themselves behind before they start.

When we hand over the compass without noticing

The nearest risk is not a machine rebelling against us. It is much simpler: that we stop thinking because something else is thinking for us.

In a study published in Radiology in 2023, 27 radiologists read mammograms alongside suggestions from an AI system. When the suggestion was correct, less experienced readers got it right about 80% of the time. When the suggestion was wrong, their accuracy dropped to 19.8% [5]. Even the most experienced readers fell from 82.3% to 45.5% [5]. Researchers call this "automation bias": we tend to believe the machine because it is a machine.

In a study by Microsoft Research presented at CHI 2025, 319 knowledge workers shared 936 examples from their daily work with generative AI [6]. The finding is striking: the more people trusted the tool, the less critical thinking they applied. The more they trusted their own ability in the task, the more critical thinking they applied [6].

In scouting, we know this lesson well. A compass doesn't walk for you. It tells you where north is, but you decide where to go, when to stop and when to turn back. A scout who follows the device without looking up at the trail can reach the edge of a valley feeling completely sure of himself.

Where is the line between empowerment and loss of control?

Empowerment is when the tool does what we couldn't, while we understand what it is doing and can stop it. Loss of control is when we can no longer understand, review, or say "no."

The problem is that the shift from one to the other doesn't happen in a single moment. It happens gradually, one small decision at a time.

The UN report Governing AI for Humanity acknowledges that no one today understands all of these systems' inner workings well enough to fully control their outputs or predict their evolution [8]. It also points to another gap: of seven prominent international AI governance initiatives outside the UN, only seven countries take part in all of them, while 118 countries take part in none, mostly in the Global South [8]. In other words, most of the world, our region included, is not at the table when the rules are drawn.

The International AI Safety Report, published in January 2025, brought together around 100 experts, with an advisory panel nominated by 30 countries plus the UN, the EU and the OECD [1]. The very existence of this effort says something: the question is no longer marginal.

On the legal side, the EU AI Act entered into force on 1 August 2024. It requires "appropriate human oversight" for high-risk systems in areas such as education and employment [9].

From our experience, we draw the line with these practical rules:

If you can't explain the decision, don't rely on it. A tool that gives you an answer without a reason is asking for blind trust.
A decision that affects a person passes through a person. Admission, evaluation, discipline and treatment are not left to the machine alone.
Compute with logic, explain with AI. Whatever can be calculated with clear rules should not be left to a probabilistic model.
Always keep the off switch. A system you can't stop or reverse is not a tool. It's a commitment.
Watch your own skills. If you notice you can no longer do the task without the tool, that's a signal worth your attention.

What building our own tools has taught us

We didn't take these rules from books. We wrote them because we ran into them while building.

When we built Kashef, our security auditor, the easy option was to ask the model to give a website a score. We didn't. The score is computed with fixed logic that any person can check step by step. The AI's role is to explain the result in plain language, not to decide it. If the model gets the explanation wrong, the number is still correct and verifiable.

In LinkGuard, we decided not to stop at "this link is dangerous." The user always sees: why this verdict? Which signals did we find? That way people learn to see for themselves, and need the tool a little less next time.

And with Najem, the Telegram assistant serving more than 1,350 scout leaders, we learned that a good tool saves time but doesn't take the leader's place. It reminds, gathers and suggests. But judging an educational moment with a young scout at camp stays with a human who knows him.

What we can do now

There is a clear gap between experts and the public. Pew found in 2025 that 56% of AI experts expect AI to have a positive impact on the United States over the next 20 years, compared with only 17% of the public [7]. Yet both sides agree on one thing: around 55% of the public and 57% of experts want more control over how AI is used in their lives [7].

That shared wish is the starting point. And control isn't granted. It is practiced.

As individuals, we can read the answer before we copy it, and ask "how do you know?" before asking "what's the answer?". We can keep at least one skill we do with our own hands every week.

As educators and leaders, we can teach young people to use the tool and to challenge it at the same time. A simple activity: give the team an answer from a language model and ask them to find one mistake in it. This builds the muscle of healthy doubt. Another: ask them to solve the problem themselves first, then compare their answer with the machine's and discuss the difference. The tool becomes a colleague to review, not an authority to obey.

As builders, we can design systems that explain themselves, keep a human in the loop, and separate what is computed from what is suggested. Transparency isn't an extra feature. It's a condition for trust.

Our region doesn't have to be a consumer only. Every school, association or small team that builds a tool with clear values adds a voice to a table that is still being set.

Smart models will draw more accurate maps every year. That much is almost certain. But a map, however accurate, doesn't choose the destination.

The tool draws the map. We hold the compass.

Human First: How Do We Make AI a Multiplier of Human Impact, Not a Replacement for It?

Moustafa hamwi — Sat, 24 Oct 2026 20:42:08 +0000

Originally published at scouthub.dev.

It's Thursday evening. A scout leader gets home from work, tired, with an activity to run for her troop on Friday morning. A few years ago, she would have spent two hours hunting for an idea and writing a plan. Tonight she opens Najem on Telegram and types two lines: the kids' ages, how many there are, and where they'll meet. A few minutes later, a full draft plan arrives.

That isn't the end of the story. She knows one of the boys is shy and doesn't enjoy competitive games. She knows the yard is smaller this week. So she cuts a section, changes a game, and gives that boy a special role. Then she goes to bed early and shows up the next morning with energy to spare.

Who did the work? The tool wrote the draft. She made the decisions, and she is the one who will stand in front of the kids and lead them. That is exactly what we mean by "human first."

The tools arrived before we were ready

AI is no longer a side experiment. Stanford's AI Index reports that 78% of organizations said they used AI in 2024, up from 55% the year before [1].

Individuals moved even faster than their organizations. A Microsoft and LinkedIn survey found that 75% of knowledge workers use generative AI at work, and 78% of those users bring their own tools without waiting for a decision from above [2].

The number that stops us is this one: only 39% of people who use AI at work have received any training on it from their employer [2].

In other words, the tool came in, but the thinking about how to use it didn't come with it. We see the same thing in associations and volunteer groups. One volunteer uses a tool to write a post, another to translate a message, and nobody has asked: what do we actually want from these tools? And what do we refuse to hand over to them?

There's also a quieter risk than mistakes. Community work runs on meeting in person. Leaders sit together, swap ideas, disagree and then agree. That's where team spirit is built, and where newcomers learn from veterans without even noticing.

When every leader gets their plan alone from their own screen, we can lose that gathering without realizing it. The plan got faster, but the team drifted further apart. So it isn't enough to ask whether a tool is accurate. We also have to ask: does it bring us together, or pull us apart?

What the research actually says

One of the clearest studies here is "Generative AI at Work." The researchers followed 5,179 customer support agents who were given an AI assistant. The number of issues they resolved per hour rose by 14% on average [3].

The more important detail is that the gains were uneven. New and less-skilled agents improved by 34%, while the effect on experienced agents was minimal [3].

We read that through a scouting lens. The new leader is the one who needs support most. They don't yet have a store of games and ideas, or experience writing a report. A good tool can save them years of trial and error. It does not replace a senior leader who walks alongside them and teaches them.

The second study draws a clear line. Researchers from Harvard and colleagues tested more than 700 consultants at Boston Consulting Group and called the result the "jagged frontier" [4].

On tasks that AI handles well, consultants performed roughly 38% to 42.5% better than those working without it. But on a task just outside the tool's ability, those who relied on it did worse, by 19 percentage points on average [4].

The lesson is simple. The tool is excellent at some things and weak at others that look very close. And it won't tell you when it's wrong. The person who knows the difference is the one who understands the context. In our work, the context is the shy child, the worried parents, and the neighborhood we serve.

Jobs are changing, and leadership is gaining value

People often ask us: will AI take people's places?

The International Labour Organization published an updated index in 2025 that answers calmly. About one in four workers worldwide is in an occupation with some exposure to generative AI. But only 3.3% of jobs fall into the highest-exposure category [5].

The ILO's conclusion is that the most likely effect is a change in what jobs look like, not their replacement, because most occupations are made of tasks that still need a human [5].

The World Economic Forum paints a similar picture. Its Future of Jobs Report 2025 expects about 170 million new jobs to be created by 2030, against roughly 92 million displaced. Employers also expect about 39% of today's skills to change over the same period [6].

What stands out is that the most in-demand skills in the report are not only technical. They include analytical thinking, resilience and flexibility, and leadership [6].

So investing in leaders is not a luxury in the age of machines. It is exactly what will stay scarce and valued.

In scouting and volunteer work, this is even clearer. Nobody joins a troop or an association because it has a smart tool. People join because there's a leader who knows their name, notices when they're absent, and is glad when they grow. Any technology that doesn't strengthen that leader, however advanced, stays outside the heart of the work.

What we learned from Najem and Smart Scout Manager

When we built Najem, our first question was: what wears a leader out without adding anything to their relationship with the kids? The answer was obvious. Writing plans from scratch.

Today Najem has generated more than 5,500 activity plans for more than 1,350 leaders in over 12 Arab countries, free of charge. But the number we're proudest of isn't the plan count. It's the hours that went back to leaders to spend with their troops.

Every plan carries the same rule: Najem suggests, the leader decides. The plan is a draft, not an instruction.

We always encourage leaders not to use the plan as is. The best thing to do with it is to bring it to the leaders' meeting, talk it through with colleagues, and let everyone add their own touch. Then the draft becomes a reason to talk, not a substitute for talking. That's what we want: a tool that opens the conversation instead of closing it.

We also believe a tool's limits should be clear. When it comes to children's safety or a specific young person's situation, we don't want AI acting like an expert. Its place there is to remind the leader to go back to their supervisor and to the child's family. Technology that knows when to step back protects people better than technology that answers everything.

Smart Scout Manager (SSM) is built on the same idea. Troops run on paperwork: meeting minutes, periodic reports, attendance records. It's necessary work, but it eats a leader's time. So we let AI write the drafts, and the leader reviews and signs. The platform's slogan sums up the goal: "Run your troop with ease, and make time for the youth."

Across 2025 and 2026, the lab's projects together reached about 17,000 people. What that taught us is that the most important decision isn't what to automate. It's what to deliberately refuse to automate.

We don't automate a conversation with a child who is having a hard day. Or assessing a young person's progress. Or resolving a disagreement between two leaders. Or calling worried parents. Those are the moments that build trust, and trust can't be written as a draft.

Before you adopt any tool

We're not asking any association or team to reject technology, or to chase it. We only suggest pausing to ask three questions before adopting an AI tool:

Where will the saved time go? If it doesn't go to people, to the youth, the volunteers, the communities you serve, you may not need the tool.
Who makes the final call? The answer should always be a person who knows the context and carries the responsibility.
Will it make our leaders stronger, or more dependent? A good tool teaches a new leader while it helps them, instead of keeping them reliant on it forever.

If your answers are clear, the tool is probably in the right place.

The technology we want is the kind that sends the leader back out to the field, among their scouts, not the kind that keeps them in front of a screen.