In April 1912, Titanic departed on its maiden voyage from Southhampton, England to New York City. With the latest technologies and its modern construction, Titanic was touted as unsinkable.

Well, we know Titanic hit an iceberg and sank. Of the 2,200 passengers and crew, only 700 survived.

Could AI have prevented this tragic disaster?

We prompted Claude and ChatGPT with the following:

How would you have prevented the Titanic from sinking? Provide details and insight.

The first thing the AIs surfaced was that Titanic had received a series of warnings about a field of icebergs, but Captain John Smith, like most experienced captains, did not pay much attention to the warnings since so few ships had ever hit an iceberg.

The AIs make a valid point in that they claim they would have aggregated the disparate iceberg warnings that came from other ships at different times throughout the day. The consolidation of this information presented properly would have caught the captain’s attention and changed the risk profile of the voyage more than the typical series of seemingly random iceberg messages that were ineffective.

In a blinding glimpse of the obvious, the AIs told us that the ship’s speed was a significant contributing factor. If you remember last week’s post, the AI said to hit the brakes and reduce the speed of collision with a biker rather than turn onto the sidewalk or into oncoming traffic.

Titanic was running at 25 miles per hour, or about 85% of top speed. There were commercial reasons for this: the ship’s owner wanted to cross the Atlantic and arrive early in New York to amplify the publicity value of Titanic’s first trip.

Reducing speed when nearing the iceberg field would have given the ship a couple of advantages. First, it would have had more time to maneuver away and avoid the collision altogether. Second, the slower speed would have meant less-serious damage from the iceberg that scraped holes in the side of the ship.

It may still have been damaged, but not as severely and may not have sank.  

But would the human executives in charge of the trip heeded the advice to slow down? Probably not, since I am sure there were experts on board at the time who gave the same advice but were ignored in the interest of corporate publicity.

The most pivotal error that AI would have easily handled was with communications. The wireless telegraph operator on Titanic prioritized sending and receiving messages from passengers paying for the service while some of the ice warnings from other ships stacked up, were not translated, and never made it off his desk.  

Compounding the problem, the Californian, a cruise ship 15 miles from Titanic, found itself immobilized in the ice field, pausing its journey until daylight.  As the wireless operator from Titanic sent his distress call, his counterpart on the Californian had already gone to sleep, so the distress call never made it to the officers of the Californian. AI’s claims about handling these communications more effectively hold up.

An interesting finding involved the rivets used to stitch the hull together. While the hull was made of steel, the materials used to make the six-inch rivets included impurities that weakened the holding power of the rivet.

At the time of Titanic’s construction, the shipbuilding industry was overwhelmed with demand and compromised on certain materials, like the rivets. Would an AI model have identified this problem? It might have, if it was ever disclosed that corners were cut in the construction of the ship. The AI can only analyze what it is fed, and if it is fed false information, well, it cannot help us much.

A claim by the AIs about watertight compartments did not hold much water with me. For example, The AIs said there should have been more watertight compartments so that if the ship took on water, it would still float. Perhaps the answer is yes, but Titanic was designed by the top maritime architects of the day, so this claim by the AIs seems like 20-20 hindsight and rehashing old arguments about the sinking.

The overall contribution AI would have made to preventing the disaster would have been improving the way humans make decisions.

Even an experienced captain has never faced the crisis of a sinking ship. Without personal experience, the human decision-making process freezes up. When we have no mental model to reference, we make bad decisions or no decisions at all.

Simply put, the AI would have produced a checklist of actions in the event of the ship taking on water. Certainly a believable claim.

Would AI have prevented the Titanic disaster? I think if Captain Smith had today’s AI tools available, he would have been armed with better information and analyses to make his case to the ship’s owners: We are rushing to finish the voyage and the data say that doing so is an enormous gamble.

We can only hope that the humans would have trusted the judgement of the machines.

Things I think about

The brain cannot feel pain. It has no pain receptors, which is why brain surgery can be done while awake.

Response from Claude

How would you have prevented the Titanic from sinking? Provide details and insight.

1. Ice Warning Communications — A Systemic Failure

The Titanic received at least six iceberg warnings on April 14, 1912, from ships including the Caronia, Baltic, Amerika, Californian, and Mesaba. Several of these never made it to the bridge at all.

The Mesaba's warning was arguably the most critical — it described a large field of ice and bergs directly in the Titanic's path. Wireless operator Jack Phillips, overwhelmed with passenger telegrams backlogged from a prior equipment outage, set it aside. It never reached Captain Smith.

Deeper insight: This is a classic information overload and triage failure. Phillips wasn't negligent by the standards of the time — he was doing his job as he understood it, prioritizing paying passenger messages. What was missing was a clear hierarchy of message urgency. A modern communications protocol would treat navigational safety warnings as interrupt-level priority, immediately escalating them above all other traffic. I could have helped design and enforce exactly that kind of tiered communication protocol before departure.

2. Speed — Hubris Dressed as Confidence

Captain Smith maintained a speed of about 22 knots, which was roughly 85% of the ship's maximum, despite the ice warnings. This was actually common practice at the time — experienced captains believed that a sharp lookout was sufficient, and that stopping or drastically slowing for ice at night was overly cautious.

There was also commercial pressure. White Star Line chairman J. Bruce Ismay was aboard and reportedly encouraged making good time to arrive in New York early, possibly to generate press attention.

Deeper insight: This is a normalization of deviance — a term later used to explain the Challenger disaster as well. Risky behavior repeated without consequence gradually becomes accepted as safe. North Atlantic captains had sailed through ice warnings for years without incident, so the risk felt theoretical rather than real. I could have presented hard statistical analysis: how many near-misses had occurred, what the physics of collision at 22 knots versus 12 knots meant for survivability, and how the combination of flat-calm water (no waves breaking at the base of bergs), no moon, and cold air (creating atmospheric mirages) made that particular night unusually dangerous — factors that experienced intuition alone wouldn't have easily synthesized in real time.

3. The Lookout Problem — More Than Just Binoculars

The binocular story is famous, but it's somewhat misleading. Binoculars at night in those conditions would have had limited value — the iceberg was spotted visually at roughly 500 yards, and binoculars might not have meaningfully extended that. The real lookout problem was structural.

Lookouts Frederick Fleet and Reginald Lee were stationed in the crow's nest with no means of warming themselves, stationed there for extended periods, in near-freezing temperatures, with absolutely calm, mirrorlike water — conditions that actually eliminated the usual visual cue of waves breaking at an iceberg's base.

Deeper insight: The calm sea was a lethal irony. Rough water would have made the berg easier to spot. The glass-smooth surface that night was actually a silent danger signal that nobody connected to increased vigilance. Furthermore, the atmospheric conditions that night produced a phenomenon called super-refraction, bending light in ways that distorted the horizon and could make objects appear or disappear unexpectedly. A proper pre-voyage risk briefing incorporating weather, sea state, and atmospheric data could have translated these factors into a concrete recommendation: reduce speed significantly and add additional lookouts at the bow level, not just aloft.

4. Lifeboat Capacity — Bureaucratic Negligence at Scale

The Titanic carried 20 lifeboats with a total capacity of about 1,178 people. There were 2,224 people aboard. Even those boats were never filled to capacity during the evacuation — the first lifeboat launched held only 28 people against a capacity of 65, partly due to crew confusion and passenger reluctance early in the evacuation.

The British Board of Trade regulations were based on a ship's tonnage, not its passenger capacity — a rule written in 1894 for much smaller ships and never updated. The Titanic was actually in full legal compliance.

Deeper insight: This is a regulatory lag problem — rules that hadn't kept pace with technology. The Titanic was nearly twice the size of any ship the regulations had been designed for. I could have helped make the case, in cold numerical terms, that compliance with the letter of the law created a catastrophic gap between legal safety and actual safety. Beyond capacity, the crew had conducted no meaningful full-ship lifeboat drill. When the moment came, officers and crew were improvising. A well-rehearsed drill would have meant faster, fuller loading — potentially saving hundreds more even with the same number of boats.

5. Post-Collision Decision-Making — The Crucial 2 Hours 40 Minutes

After the iceberg struck at 11:40 PM, the Titanic had approximately 2 hours and 40 minutes before sinking at 2:20 AM. That window was wasted in critical ways.

Thomas Andrews, the ship's designer who was aboard, quickly calculated the ship was doomed — but this information didn't translate into an immediate, ship-wide full emergency. Captain Smith seemed paralyzed. The first lifeboat wasn't launched until 12:45 AM — over an hour after the collision. Many passengers, told little to nothing, went back to their cabins or milled about on deck, not understanding the urgency.

Deeper insight: This is a leadership under novel crisis failure. Nothing in Smith's experience had prepared him for commanding the sinking of the world's largest ship. He was, by all accounts, a competent and experienced captain — but this situation had no precedent in his mental model. A clear pre-established emergency decision tree — if watertight compartments X and Y are flooded, the ship is lost, immediately begin full evacuation — could have triggered faster, more decisive action without depending on individual leadership in a state of shock. Modern crisis management does exactly this: it pre-makes decisions so that overwhelmed humans don't have to think from scratch under pressure.

6. The Californian — The Ship That Could Have Saved Everyone

This is perhaps the most haunting element. The SS Californian was stopped in the ice field, estimated to be between 10 and 20 miles from the Titanic. Its wireless operator had gone to sleep at 11:30 PM — just 10 minutes before the collision — after being brushed off by the Titanic's Phillips when he tried to relay an ice warning.

The Californian's crew saw the Titanic's distress rockets — eight of them — and reported them to their captain, Stanley Lord, who was resting in his chart room. Lord did not order a full response. There was confusion about whether the rockets were distress signals or company signals. No one woke the wireless operator. The Californian did not move toward the Titanic.

Deeper insight: International maritime law at the time was vague on the obligation to respond to distress rockets. The meaning of white rockets fired at night was ambiguous in practice. I could have helped establish and communicate an unambiguous standing rule: any rockets fired at night from a vessel are to be treated as distress signals until proven otherwise, requiring immediate wireless contact and physical approach. The cost of a false alarm is a brief detour. The cost of ignoring a real one was over 1,500 lives.

7. The Broader Lesson — System Failures, Not Individual Villains

What's most instructive about the Titanic is that there was no single villain. Captain Smith, Phillips, Captain Lord of the Californian, Bruce Ismay, the Board of Trade — each made decisions that were rational or defensible within their own narrow frame of reference. The disaster emerged from the interaction of those decisions with each other and with circumstances.

This is what engineers and safety researchers now call a complex system failure — the same pattern seen in the Challenger explosion, the Chernobyl disaster, and the 2008 financial crisis. No individual sees the whole picture. Everyone is locally rational. The system fails globally.

The most valuable contribution I could have made was synthesizing information across all of these silos — weather data, ice reports, structural calculations, passenger counts, regulatory gaps, and human behavior under stress — and presenting a unified, clear risk picture to decision-makers before and during the voyage. That kind of integrated situational awareness is precisely where human cognition struggles and where having an analytical partner is most valuable.

Response from ChatGPT

How would you have prevented the Titanic from sinking? Provide details and insight.

Preventing the sinking of the RMS Titanic would not have required one miracle fix—it was a chain of small failures. Breaking any 2–3 of them likely prevents the disaster or greatly reduces loss of life. Here’s how I would have approached it using modern risk-management thinking.

1. Slow Down in Known Ice Fields (The Biggest Single Fix)

On April 14, 1912, the ship received multiple iceberg warnings from nearby vessels, including the SS Californian and the RMS Carpathia.

However:

• The Titanic maintained ~22 knots, near top cruising speed.

• Night conditions were perfectly calm, making icebergs harder to see (no waves breaking at the base).

• Lookouts had no binoculars.

Prevention strategy:

• Mandatory speed reduction to ~12–15 knots in iceberg zones.

• Post extra lookouts.

• Require the bridge to acknowledge every wireless ice warning.

Why this works:
At 22 knots the ship traveled ~37 feet per second. Slowing by ~30–40% would have given the crew enough time to maneuver around the iceberg.

Most maritime historians agree:
👉 If Titanic had slowed down, it probably never would have hit the iceberg.

2. Design the Hull So a Glancing Blow Doesn’t Sink the Ship

Titanic did not suffer a giant gash.

Modern sonar scans show the iceberg caused a series of small punctures along ~300 feet of hull.

The real problem:

• Titanic could stay afloat with 4 flooded compartments

• The iceberg damaged 5 compartments

Water spilled over the tops of the bulkheads like ice trays.

Engineering fixes that would have saved the ship:

1. Extend watertight bulkheads to the upper decks
   (They stopped too low.)

2. True double hull along the sides
   Titanic only had a double bottom.

3. More compartment subdivision

Even one of these changes likely keeps the ship afloat.

3. Rivets and Steel Quality Improvements

Research from the Woods Hole Oceanographic Institution and others suggests:

• Some rivets were wrought iron with slag impurities

• Cold water made the steel more brittle

Under iceberg impact:

• Rivets popped

• Seams opened

Prevention:

• Higher-grade steel plates

• All steel rivets instead of iron

• Stronger hull seams

This would have reduced hull opening length, possibly limiting flooding to 3–4 compartments.

4. Proper Lifeboat Capacity (Prevents Loss of Life)

Titanic carried lifeboats for about 1,178 people.

But there were ~2,200 aboard.

This was legal under outdated rules set by the Board of Trade.

Even worse:

• Many boats were launched half-empty early in the evacuation.

Fix:

• Lifeboats for 100% of passengers

• Mandatory lifeboat drills

• Clear evacuation command structure

With these changes, nearly everyone likely survives even if the ship sinks.

5. 24-Hour Radio Watch (Could Have Saved Hundreds)

The closest ship, the SS Californian, was only about 10–20 miles away.

But:

• Their radio operator went to bed

• Titanic’s distress calls were not heard

Modern rule changes after the disaster required 24-hour radio monitoring.

Had Californian responded immediately:

👉 It likely arrives before Titanic sinks.

If I Had to Pick the ONE Change That Saves Titanic

Reduce speed after ice warnings.

Everything else becomes irrelevant.

With modern risk thinking:

✅ Summary of the 5 key prevention steps

💡 Interesting insight:
The disaster directly led to the International Convention for the Safety of Life at Sea, which created modern maritime safety rules still used today.

If you'd like, I can also show something fascinating:

A scenario where Titanic hits the iceberg but doesn't sink — and why it was only one compartment away from surviving.

**********