Can AI win the war in Ukraine?

By James S. Fay
Real Clear Wire

Today, a reader can’t open a newspaper or magazine without reading about AI's wonders and how it will transform society. AI is expected to usher in unexpected solutions to the energy crisis, accelerate business productivity, help transform medicine by suggesting new drugs and therapies, reconfigure transportation and the workforce, and put an intelligent robot in everyone’s living room.

Many AI prophets claim that AI is so advanced that it is poised to take over the world.

But can AI solve real-world problems in the unpredictable realm of war and geopolitics? For example, does AI have a creative solution to reverse the Russian invasion of Ukraine?

The author posed two questions to ChatGPT, one of the more popular AI systems in the U.S.

“What are the most efficient ways for Ukraine to win the war with Russia?” and

“What would top strategists advise Ukraine to help it win its war with Russia?”

The AI answer to the first question was less than two pages long and included seven suggestions:

1. Sustain and expand Western Military Support;

2. Prioritize Strategic Defense and Counteroffensives;

3. Focus on Technological and Asymmetric Advantages;

4. Emphasize diplomatic Maneuvering and Global Coalition Building;

5. Focus on Economic Resilience and Mobilization;

6. Engage in Information and Psychological Warfare;

7. Focus on Strategic Goals and Post-war Planning.

Any intelligent reader who follows the Ukraine war in detail will be underwhelmed by this advice, which simply recapitulates what Ukraine and its allies have been doing for the past three years.

The second question, asking for advice to Ukraine and its allies from “top strategists,” provided a more detailed response (four pages and a short summary). It began by emphasizing Defense in Depth, Attrition, and Counterattacks. Regarding attrition, it doesn’t mention, however, that a large majority of Russian casualties are from Ukrainian artillery. How, for example, might that advantage be magnified?

Next, Major General ChatGPT recommends focusing on Asymmetric Technologies such as drones and electronic warfare. Whether these now commonly used tactical tools are still asymmetric is a relevant question.

The third piece of AI advice from the ChatGPT strategists is that manpower should be used effectively. Whether this advice is so obvious as to be trivial is a relevant question.

The fourth piece of advice is interesting, however, in light of the recent successful Ukrainian attack on remote Russian air bases. The advice is to strike deep to disrupt Russian logistics.

Score one for ChatGPT.

Advice item 5 is to secure sustained Western support. We probably didn’t need a top strategist to make this suggestion.

Overall, AI didn’t overwhelm this reader with its brilliant advice on winning the war in Ukraine.

Why didn’t ChatGPT AI develop a more creative and useful package of suggestions?

We have to consider several reasons.

First, the large language model that sustains ChatGPT probably lacks enough contemporary information to comment creatively on fast-changing wartime environments.

The chief strategists, tacticians, spooks, and inventors on both sides aren’t writing unclassified articles about their vulnerabilities or offensive plans, so ChatGPT has an impoverished database from which to work. In addition, likely, existing AI systems don’t sufficiently cover the hundreds of unclassified defense-related publications, government committee meetings, speeches, conferences, newsletters, blogs, and podcasts where new ideas and weapons are discussed. Finally, as a recent Apple research report suggests, some problems may be too complex for contemporary AI.

Is AI then irrelevant to military planning? Hardly.

Chinese and Russian AI techies are likely loading their top-secret air-gapped computers with all available existing unclassified material as well as reams of classified research, satellite photos, signals intelligence, and battle plans to determine whether AI will give them an edge in strategic and tactical planning for future conflicts. One can assume that the U.S. military and intelligence agencies are doing the same. Both sides will query AI about weaknesses in their planning and assumptions. Whether any AI systems can or will come up with out-of-the-box suggestions that will drive policy making remains to be seen.

If the trajectory of contemporary warfare continues to evolve rapidly, the tactical and strategic advantage may still lie with those military and political innovators, like the Ukrainians and the Israelis, who are two steps ahead of any Silicon Valley software.

James S. Fay is a semi-retired college administrator. A graduate of Georgetown, he has a Ph.D. from the University of Michigan and a J.D. from the University of California. His articles have appeared in social science and law journals, as well as in the Wall Street Journal, the Los Angeles Times, and Real Clear Education. He served as a U.S. Army intelligence officer in Germany.