Matrix Games Forums

People have started writing books with AI. My experience is that an actual novel created by an AI is usually mediocre at best (and frequently much worse). However, as a brainstorming tool, an AI can be really useful.

For example, I'm currently writing a series of scenarios about a global crisis set in 1968. I was wondering what might happen with Cuba within this larger situation. So, I asked Bing.

I assumed, based on earlier research, that the Soviets might have a SAG in the area. So, I asked Bing a series of questions about the composition of the SAG, possible missions, how missions might play out, etc.

Short version for the "too long, didn't read" generation...I'm not talking about teaching an AI to write scenarios directly for Command. I do not have that kind of programming skill, and I suspect even if I did it simply wouldn't be worth the time and effort to code, debug, and train an AI to do this. However, as a brainstorming tool, throwing out questions and getting back ideas, possible details, etc., it's nice. Definitely worth trying out. Of course, I'll still have to write the actual scenario (adding units, setting up missions, etc.) and I'll have to check a lot of the details, but I was very pleased by how my exercise turned out. I'd be interested in hearing what people think of this.

Some interesting details...

Bing quickly gave me an order of battle (Kynda, Kotlin, Riga, and Foxtrot) and several mission possibilities. Here is a quote of part of its answer:

1. Sea Denial and Interdiction

Objective: Disrupt U.S. and allied shipping routes in the Caribbean and Gulf of Mexico.

Tactics: Use the Kynda-class cruiser’s long-range SS-N-3 Shaddock missiles to threaten or sink merchant vessels, oil tankers, or lightly defended naval auxiliaries.

Impact: This could have created a temporary choke point in the Gulf, especially near the Yucatán Channel or Florida Straits.

2. Threat to the Panama Canal

Objective: Temporarily disable or threaten the Panama Canal to disrupt U.S. logistics and force redeployment of U.S. naval assets.

Feasibility: While a direct assault would be risky and likely suicidal, the SAG could:

Launch a missile strike on canal infrastructure (e.g., locks or control stations).

Conduct a diversionary feint to draw U.S. forces away from other theaters.

Deploy saboteurs via covert landing or support a submarine-launched operation

3. Coastal Bombardment or Missile Threats

Targets: U.S. installations in Puerto Rico, Guantánamo Bay, or even oil infrastructure along the Gulf Coast.

Means: The Kynda’s cruise missiles had a range of up to 450 km, allowing standoff strikes from international waters.

Psychological Effect: Demonstrating reach and resolve, potentially forcing the U.S. to divert air and naval assets from Europe.

4. Support for Cuban Defense

Role: Act as a mobile seaborne extension of Cuban coastal defense.

Function: Provide radar coverage, anti-aircraft defense, and anti-submarine screening against limited U.S. naval incursions.

5. Intelligence and Surveillance

Assets: The Riga-class frigate and accompanying support vessels could conduct ELINT/SIGINT operations.

Goal: Monitor U.S. Coast Guard and remaining Navy movements, especially around Key West, Guantánamo, and the Bahamas.

I asked a few more questions, got some additional possible details, and even a fictional contemporary news account. Again this is a direct quote:

CBS Evening News Special Report August 14, 1968 ANCHOR: Walter Cronkite

"Good evening. Tonight, the world watches the Caribbean with bated breath as a Soviet naval task force, operating from Cuban waters, has triggered the most serious maritime standoff since the Cuban Missile Crisis of 1962. The Pentagon has confirmed that a group of Soviet warships, including a guided missile cruiser, a destroyer, and a frigate, approached the Yucatán Channel earlier this week, prompting a rapid U.S. military response."

Again, I'm not sure what kind of scenarios you would get if you actually had an AI program a computer directly, and I definitely don't know if it would be worth the effort of doing all that coding and debugging and training. Certainly, for someone with my limited programming skills, doing this would be a non-starter. However, a brainstorming tool, for exploring and testing ideas, proposing details, etc., it has a lot of potential.

Thoughts on all this?

It sounds useful, especially for trying to come up with a large campaign of scenarios. I personally wouldn't use it as the sole source for ideas, but quickly coming up with my own and then weighing it against anything the AI has to say would enhance the creative process.

For folks with limited time it would be helpful in moving from the planning phase to the "production" phase of scenario building faster.

At the same time, for a campaign I'd just want to be careful and make sure that each scenario the AI comes up with is coherent. I wouldn't trust it to remember changes to an OOB at the outset of the war (I'd be wary of it proposing to put 13 of a class of cruiser in different places around the world, when only 4 of that cruiser were even built), and I also wouldn't trust it to keep track of any "canon" changes to the OOB (like a major surface combatant being sunk in a previous scenario). But as a tool for doing the more basic legwork of scenario creation, it would still be quite a boon.

Are you on discord? I'd be happy to show you where oceanic horizons sits right now. I've built it entirely with AI