AI Deepfake Scams Hit 16.6 Billion Dollars — And Your Data Isn't Safe Either

In 2024 alone, cybercrime drained more than 16.6 billion US dollars from victims, according to the FBI's IC3 annual report — and a growing share of that money was stolen with help from artificial intelligence. The AI deepfake scam of 2026 is not a horror story about the future. It is happening now, it is cheap to run, and it does not need your bank password to hurt you. It just needs a few seconds of your voice, a photo, or a document you handed over without thinking.

What Google's June 2026 fraud advisory is warning everyone about

Google published a fraud advisory in June 2026 laying out how quickly AI-powered scams are spreading. The pattern is consistent across the report: criminals are no longer sending clumsy emails full of spelling mistakes. They are cloning voices, faking video calls, and writing phishing messages so clean that people who should know better are clicking anyway. The tools that used to require a skilled team now run from a laptop, which is exactly why the numbers are climbing.

Security researchers at Vectra AI and identity-verification firm Sumsub have tracked the same shift. Sumsub's data points to deepfakes now making up around 11 percent of all fraudulent activity worldwide. A few years ago that figure was a rounding error. Today it is one in nine fraud attempts involving synthetic media that did not exist until a machine generated it.

AI fraud, in real numbers

It is easy to wave away statistics, so here are the ones that should make you sit up. These are the AI cybersecurity risks that are already costing people and companies real money:

• ✦A single deepfake video call cost the engineering firm Arup 25.6 million US dollars. Staff joined what looked like a normal meeting with senior colleagues — every face on the call was fake.
• ✦Voice cloning needs only 3 to 10 seconds of audio. A voicemail greeting or a TikTok clip is more than enough to copy how you sound.
• ✦AI-generated phishing emails get roughly four times the click rate of ones written by humans, because the machine writes faster, cleaner, and more convincingly at scale.
• ✦The World Economic Forum found that 73 percent of organisations were directly affected by cyber-enabled fraud in 2025 — nearly three in four.

WebProNews and others have reported on the Arup case as a wake-up call for big companies. But the same techniques scale down. A scammer does not need a 25 million dollar target to bother cloning a voice. A worried parent who gets a panicked call in their child's exact voice is a target too.

Why this matters to students, not just corporations

You might be reading this thinking it is a problem for banks and engineering firms. It is not only theirs. The fuel for AI fraud is data, and students hand over data constantly without realising it. Every photo, every voice note, every document you upload somewhere becomes raw material for a system you do not control.

Here is the part nobody warns you about. When you paste your assignment brief, your draft, or your personal details into a free AI chatbot to get an essay written, you are not having a private conversation. Depending on the tool and its settings, that text can be stored, reviewed, and used to train the next version of the model. Your half-finished thesis, your student ID details, the personal anecdote you wrote for a reflection paper — all of it can leave your hands and never fully come back.

Your brief should go to a person, not a language model

This is the quiet risk hiding inside the AI shortcut. People worry about Turnitin catching AI writing, and they should. But there is a second problem: the data you feed the machine to get that writing in the first place. You are trading your private work and identity for a result you cannot defend, while possibly handing a corporation a free copy of everything you typed.

At CodedByHumans, your brief goes to a real graduate. A person reads it, writes it by hand, and sends it back. It does not get swallowed into a training set. It does not become a data point. There is no model on the other end memorising your words — just a human being doing the work the way it was always meant to be done.

“Feed a chatbot your brief and you feed the machine. Feed it to us and it goes to a human who forgets it the moment the work is done.”

How to protect yourself from AI scams in 2026

You cannot stop criminals from building these tools, but you can make yourself a harder target. A few habits go a long way:

• ✦Agree on a family code word for emergencies, so a cloned voice asking for money can be checked instantly.
• ✦Be sparing with voice notes and public clips — a few seconds of audio is all a cloner needs.
• ✦Treat any urgent request for money or login details as suspicious, even if the face and voice look right.
• ✦Never paste personal information, ID numbers, or private documents into a free AI tool you do not control.
• ✦Slow down. Urgency is the scammer's favourite weapon, because it stops you from checking.

The thread running through every one of these is the same: protect your data, because your data is what makes the fraud possible. The less of yourself you scatter across machines you do not own, the smaller a target you become.

That is the whole idea behind CodedByHumans. Real graduates, zero AI, Turnitin-safe — and your work stays between you and a person, never a server farm. If you need an essay, report, dissertation or coding task done properly and privately, message us on WhatsApp for a free quote. Nothing to detect, because nothing is AI — and nothing of yours gets fed to a machine along the way.