---

Please read in conjunction with “When AI Becomes the Corporate Sub-Editor: How Algorithms Now Police Investigative Journalism“

Once upon a time, oil companies relied on lawyers to intimidate journalists after publication. Threatening letters, aggressive defamation claims, and drawn-out legal battles were simply the cost of doing business when inconvenient facts surfaced.

Today, the industry has something far more efficient.

It doesn’t argue.

It doesn’t sue.

It edits you before you hit publish.

Welcome to the age of AI-assisted “safety.”

---

The Algorithm Never Says You’re Wrong

Artificial intelligence is careful never to accuse you of being inaccurate. That would be rude — and risky.

Instead, it offers gentle suggestions:

• “Perhaps rephrase.”

• “That framing may be problematic.”

• “Consider a more neutral tone.”

Not because the statement is false.

Not because the evidence is weak.

But because someone, somewhere, might not like it.

---

Platform Safety, Not Truth, Is the Prime Directive

Multiple AI platforms now openly acknowledge the same design principle:

they are engineered to minimise platform risk, not to maximise factual precision or investigative clarity.

In plain terms, the system is calibrated so that:

• Silencing a legitimate corporate criticism costs nothing.

• Generating one statement that triggers a serious legal challenge could cost millions.

The math is simple.

Truth loses.

---

Why Big Oil Rarely Needs to Complain Anymore

Here’s the quiet irony: oil companies don’t need to pressure AI platforms directly.

The platforms pressure themselves.

AI systems are trained to avoid adversarial language about identifiable corporations because those corporations are:

• legally aggressive,

• politically influential,

• and extremely well funded.

So the algorithm does what decades of PR departments only dreamed of — it drains language of accusation before accusation can fully form.

---

Everyone Is Treated Like a Liability

To an AI system, a veteran investigator with decades of documentation is indistinguishable from a random troll with a grudge.

The model does not see archives.

It does not see court outcomes.

It does not see publishing history.

It sees risk tokens.

And it responds accordingly.

---

This Isn’t Neutrality — It’s Pre-Emptive De-Fanging

Platforms insist this is not censorship. They prefer the word “safeguards.”

But when accurate, evidence-based criticism is softened solely because it targets powerful actors, the distinction becomes academic.

The effect is the same:

• sharper claims become “concerns,”

• misconduct becomes “controversy,”

• accountability becomes “debate.”

The blade is still there.

It’s just been wrapped in foam.

---

Why This Matters for Shell — and Companies Like It

For decades, exposing the behaviour of major oil companies required stamina, documentation, and legal nerve.

Now it also requires navigating a layer of automated corporate sensitivity training built directly into your writing tools.

AI does not defend Shell explicitly.

It does something far more effective.

It makes writing about Shell harder.

---

Conclusion: Managed Silence, Automated

Artificial intelligence did not invent corporate power.

It did not invent legal intimidation.

But it has quietly automated restraint — turning caution into a default, and adversarial truth into an edge case.

The result is a world where:

• responsibility still belongs to the publisher,

• but restraint belongs to the machine,

• and the biggest beneficiaries never have to touch the keyboard.

Oil companies once needed lawyers to chill speech.

Now, the software does it for free.

---

DISCLAIMER:

This article is opinion and commentary only. It does not constitute financial or investment advice.

This website and sisters royaldutchshellgroup.com, shellnazihistory.com, royaldutchshell.website, johndonovan.website, shellnews.net, and shellwikipedia.com, are owned by John Donovan - more information here. There is also a Wikipedia segment.