How AI-washing is scamming recruiters

Let me make this crystal clear.

And I will try to do that in two statements.

At first, they might appear contradictory, but they are far from that. Two conflicting ideas can be true at the same time.

You must make up your own mind on this too, because it affects you in every way, whether you like it or not.

1: I believe clever, well-researched, thoroughly tested, and ethical AI will absorb a lot of the work currently done by recruiters. It will also be good for recruiters, candidates and clients of recruitment companies. Efficiency, accuracy and cost saving will flow, and the recruiters who have the new required skill set will thrive and be happier in their roles.

2: Much of the AI being sold to recruiters right now is flawed. It’s bad. In some cases, it is a flat-out scam. It can’t do what it promises. It’s often not even AI. Just basic automation. It’s trained on poor or incorrect data, it cannot understand context, it cannot reason, and it does not improve over time. And as for compliance, privacy and discrimination? Lord help us! It is biased, inconsistent, untested, and overpromises – but underdelivers

So, I am pleading with recruiters and recruitment owners. Run towards the tech. Make it your business to understand enough to know what to look for and what to ask.

The implications are enormous. You need to decide what to automate and what to keep human. You need to select quality AI that works in harmony with your tech stack and that improves over time. You need to grow the skillset in your business to get the best out of an AI/ Human hybrid world

Florian Fisch, a German recruiter with a keen interest in AI and real tech solutions for the recruitment industry, saw my warnings on this and wrote me a piece for The Savage Truth. So over to Florian. Connect with him and see what he is up to.

Here is the piece by Florian Fisch

***********************************************

Greg Savage said it perfectly: “Recruitment is the only industry where what you’re selling can say no.”

After three years of recruiting for BMW, I thought technology would solve the human complexities that made perfect candidates vanish. Then I moved to the tech side, building recruitment AI. What I discovered shocked me.

Right now, agencies globally are paying premium prices for “AI-powered” tools using technology from the 1990s.

The AI-Washing Playbook: 7 Tricks They Use

1. The “Powered by AI” Badge
Claim: “AI parsing understands resumes like recruiters.”
Reality: Regular expressions from 1987.
Test: If it can’t match “Led development teams” to “Team Leader”—it’s not AI.

2. The Machine Learning Mystique
Claim: “ML algorithms continuously improve.”
Reality: A/B testing keywords. Zero learning.
Test: Real ML shows improvement curves. Fake ML fluctuates randomly between 75% and 80%.

3. The Cognitive Engine Con
Claim: “Cognitive technology thinks like recruiters.”
Reality: IF-THEN statements from Programming 101.
Test: Search “Team Leader” when CV says, “Led teams.” Cognitive AI understands this. Keyword matching doesn’t.

4. The NLP Nonsense
Claim: “Advanced NLP understands context.”
Reality: Basic stemming from 1968.
Test: Try “Didn’t manage teams but led projects.” Real NLP catches negation. Fake NLP sees “manage teams.”

5. The Predictive Analytics Pretence
Claim: “AI predicts candidate success.”
Reality: Linear regression on experience and keywords.
Test: Ask about career changers. Real AI grasps transferable skills. Fake prediction fails.

6. The Deep Learning Deception
Claim: “Deep neural networks for unprecedented accuracy.”
Reality: Database queries are called “deep” for marketing.
Test: Ask about architecture. Real: layers, neurons, backpropagation. Fake: “proprietary algorithms.”

7. The 99% Accuracy Absurdity
Claim: “Industry-leading 99% accuracy.”
Reality: Measuring meaningless metrics like email extraction.
Test: Ask what they’re measuring. Watch them squirm.

What Real AI Actually Looks Like

True AI understands context—”Python developer” implies Django knowledge even if unlisted. It grasps implications—”startup experience” signals comfort with ambiguity. It recognises relationships—”200% sales increase” means different things at startups versus Fortune 500 companies.

Real AI achieves 99%+ parsing accuracy. Keyword matching delivers 78%. That 22% gap? Those are qualified candidates you’ll never see.

The Real Cost

When you use fake AI:

• 22% of qualified candidates are invisible in your database
• Your team manually reviews everything because they don’t trust the results
• Competitors find talent you miss
• You pay $200-400 monthly for $20 technology

Five Questions to Spot Real AI

1. “Show me accuracy metrics on real CVs from my market.”
2. “How do you handle implied but unstated skills?”
3. “What happens with different terminology for the same role?”
4. “Can you parse scanned PDFs?”
5. “How does your system improve—specifically?”

The Path Forward.

The recruitment industry deserves better than repackaged 1990s automation at 2025 AI prices.

Test your tools using these questions today. Calculate real costs: wasted time, missed candidates, inflated fees. Demand transparency: accuracy metrics, demonstrations with messy data, and technical explanations.

The more recruiters demand proof over promises, the faster AI-washing exits the market.

Vendors won’t thank me for this. Some threatened legal action before. But after watching agencies pay thousands for keyword matching while believing they bought AI, someone had to speak up.

The question isn’t whether we need technology—we absolutely do.

The question is: Will we keep paying AI prices for 1990s automation, or demand actual intelligence?

****************

Florian Fisch is a former BMW recruiter turned recruitment technology expert, specialising in exposing AI-washing and building genuine AI solutions for the recruitment industry.

Connect with him on LinkedIn 

Reach out to chat about any of this

************************************************************************************************************************************************************************************

Now is the time to develop your team

The Savage Recruitment Academy

Now, one all-inclusive subscription level

Quarterly subscription as low as $59 per quarter when signing up multiple users

Includes 12 learning pathways. 200 hours of video training. Smart “Ask Greg” AI Assistant to ask advice from, clients and candidate personality profiling, testing, and much more!

And new “Agency size” pricing, so you can get logins for the entire team, at very reasonable rates

Home page

Course Content

Pricing

*********************************************************