Hello all!

In our previous OpenDevs, we already received a lot of feedback on the AI in Humankind, but we'd still like for you to continue sharing your feedback on it.

Here are some questions we'd like you to consider in your feedback:

• Does the AI play in a way that feels consistent with their Archetypes and Biases seen in the diplomacy screen?
• How did the AI evolve throughout the game? Did it feel equally challenging at the start and end of your game? Does increasing the difficulty increase the challenge as much as you expect?
• Does the AI commit any grave strategic missteps? What is the AI doing and what do you think it should do? (Please provide saves if you can.)
• How does the AI perform in battle? Does it make any obvious mistakes it should avoid? What would you have done in that situation? (Please provide a save if you can)

Please let us know what you think!

I just finished my first game of the beta, and I gotta say, the Ai just feels so much better to me, some of their moves felt almost human.

The AI did two very interesting thing to me during my game:

1.  During a war I was fighting over a territory with one of my neighbors they burned down my outpost then made one themselves, I then burned down their outpost, and before I could make a new one (bear in mind this is the very next turn), someone else built an outpost, stealing this fought over land.

2.  I had a vassal who did not like me.  The turn I declared war with a big power he declared independence.

Has the AI done anything interesting in anyone else’s games?

Just finished my first game, so I cannot comment on trands over the grand scope, but two things I noticed:

1. The battle AI is very good. It consistently makes what is clearly a good (and most fo the time the best) moves, and punished mistakes. It hunts every single terrain modified it can, and manages to use the unique abilities of most units well.

2. The biggerst mistake I have seen is that the AI is bad at finding a general tech strategy, and fails to adapt once I have an advantage in military tech. Despite reaching far into the Contemperary era (which remember, I don't even get to enter) the AI was still fielding Spearmen and Horsemen against my Line Infantry. My gut feeling is that unless it is their emblematic unit, they don't really build advanced armies. This makes them vulnerable to a late game backstab, where I stay friends most of the game (to avoid the early game where we have comperable tech) and then attack them once I get gunpoweder units.

I think the AI under-commits to religion. Religion is strong, but the AI doesn't seem to value it. Example (granted, this was Nation-level AI):

Note there are only 3 religions in the game, and the top one has zero followers. I didn't go hard into faith bonuses, indeed, the only thing I built with a faith bonus was the Scriptoriums from the Franks, and I chose no policies to boost my faith gain, I went heavy progress. The only thing that really tilted me towards Faith was choosing Polytheism to start with, since I wanted a head start on conversion.

Question regarding religion. Does changing the name of your religion confer any benefit? What I've read about it indicates that all you get is a different building. It's sort of unclear why to even bother if all you get is a cosmetic change.

AOM wrote:

Question regarding religion. Does changing the name of your religion confer any benefit? What I've read about it indicates that all you get is a different building. It's sort of unclear why to even bother if all you get is a cosmetic change.

Purely cosmetic. But it also doesn't cost you anything.

Previous games I was pretty regularly beating AI on civilization now it's very hard.

I am not sure what kind of bonuses AI gets but it does feel pretty uphill task, I checked graphs at the end of the game and I was barely getting any research while AI got 400 on turn 50 despite population being basically the same, One AI's production was 800 at that point. I honestly not sure how it was possible,

I am just curious what kind of bonuses AI gets at this difficulty.

• Does the AI play in a way that feels consistent with their Archetypes and Biases seen in the diplomacy screen?

• How did the AI evolve throughout the game? Did it feel equally challenging at the start and end of your game? Does increasing the difficulty increase the challenge as much as you expect?

• Does the AI commit any grave strategic missteps? What is the AI doing and what do you think it should do? (Please provide saves if you can.)

• How does the AI perform in battle? Does it make any obvious mistakes it should avoid? What would you have done in that situation? (Please provide a save if you can)

I just wish the AI could be better without cheating so much, it's insane how huge the bonus on Humankind difficulty is.

I also found the combat AI quite decent in most situations. It only looked a bit confused when unable to reach my units.

The difficulty increase from normal to empire is definitely steep. On normal difficulty, it's a walk in the park with no opposition from the AI whatsoever. On empire, the AI seems to get some massive tech advantage among other things (It researches its first 3 technologies in 3 or 4 turns). Does it get free techs? Also its much more aggressive (probably because it will actually find itself in a stronger situation) and fields quite large armies. Looking at the district that the AI has build however, it doesn't seem like the AI should have that high yields, since it only has a handful of districts. Personally, I think the AI boni are a bit too strong early on for the difficulty just one above normal.

AOM wrote:

Question regarding religion. Does changing the name of your religion confer any benefit? What I've read about it indicates that all you get is a different building. It's sort of unclear why to even bother if all you get is a cosmetic change.

Changing to a historical religion may occasionally trigger related events (for example, there is a midrash vs kabbala for Judaism), but otherwise, purely cosmetic.

MorshuHimself wrote:

I've been fighting the Persian AI for a while now, and EVERY SINGLE ARMY THEY HAVE are filled with Immortals. No cavalry, no range, nothing. I haven't even seen scouts floating around from them. I think with this area I've been fighting in, a canyon with a narrow pass, they would have benefitted greatly with some archers, but they elect to build exclusively Immortals. They are good units mind you, but I would think to have some ranged support.

I’ve noticed the AI is really good at building their unique units, and nothing else. . . 

MorshuHimself wrote:

I've been fighting the Persian AI for a while now, and EVERY SINGLE ARMY THEY HAVE are filled with Immortals. No cavalry, no range, nothing. I haven't even seen scouts floating around from them. I think with this area I've been fighting in, a canyon with a narrow pass, they would have benefitted greatly with some archers, but they elect to build exclusively Immortals. They are good units mind you, but I would think to have some ranged support.

The significantly lower combat strength of archers makes them not a good fit, even if there are some situation where they could have been good.
Immortals are very powerful and will blast through anything with archers in the long run.
Even with better ranged units (Markabata) and phalanx I had very tough fights against those immortals.

I think in the early game the AI is doing somehow well. Especially the problem of Victor that the AI would declare an unjust war without having a military to support it seams to be less of an issue now.

But in the late game the AI is getting a bit weaker compared to a human. So once you survived the early game it becomes less challenging to deal with the AI. I think the reason for this is that the AI is not expanding in the mid and late game. The AI does not look for other continents and other land to settle. This means that the other continent in this beta was left for human players only. The AI should make this a priority to research the required technology and sending troops there to colonize. I think that would make the AI much more competitive in the later parts of the game.

Otherwise, I think the AI is transcending more often than it should. Since the AI is not expanding, it does can't really make use of the EQ in new territories. So the main benefit goes away. I think especially on higher difficulties the AI should be better at picking cultures than benefit their long-term agenda and that synergize well with each other. I felt like AIs often stay within on affinity, which I think is not always the best idea. The question of how to pick cultures is generally a really hard one and I appreciate that it is harder to program an AI to do this smart, but there are some obvious synergies that the AI should plan for and that it currently does not seem to make use of.

Another idea how to solve the scaling problem is to maybe slightly increase their buffs for each era. This would make up for the lack of strategy, but it would certainly be more fun if  the AI was actually smart.

Mostly agree with ritchiaro.

I think AI is generally decen,t I did not see any glaring mistakes, but this beta AI does not look new world and it's indeed not getting enough fame.

In my game I had a lot of difficulties initially but after some point I was easily able to outscore AI in mostly peaceful game,  As an idea AI should estimate how much fame it could potentially get in next x turn vs advancing right away. It's almost always better to wait to get a few high scoring stars than advance immediatly.

Fascinating. The AI somehow managed to obliterate another civilization! I've never been able to do so myself, always ran out of time then just vassalized them, but this was interesting.

AI seems better now. Some areas they could improve are the following:

1. When sieging them, you can park your units just outside without attacking and they will often attack you, effectively nullifying their walls.

2. They don't upgrade units to the current era fast enough

3. I brought in some u-boats to a siege (which can't attack land units), and the AI focussed them instead of the one unit I had that could actually attack them. This is a pretty edge case scenario, but I thought I should mention it anyways.

4. It would be really cool to see the AI make a coaltiion against you if you've vassalized a number of them.

I agree with the sentiment that the AI is strong early and falters mid to late game. I also think the difficulties should be harder and better tuned, because I should not be playing my second or third game kind of casually and beating the Civilization difficulty fairly easily (did not play much of the earlier betas, did not touch Victor opendev). The highest difficulties should present a challenge where I should be planning my turns one by one while already having a really strong grasp of the systems. Although it may be possible that I'm not feeling the difficulty as much because this map seed kind of favors the player because there's a lot of open space to the right and I did fast settle to the natural wonders on the south to immediately get the influence boost while boxing my closest AI neighbor in. The systems in this game are great but require a lot more tuning to present an interesting and continuous challenge. I do agree with other posts in this forum that mention AI cities on higher difficulties get too much of a stability boost to be useful to human players, higher difficulties shouldn't change gameplay too substantially that causes things like capturing AI cities no longer make sense. 

On a different point, not sure what the design intent of allied or vassalized AI are, or what AI expects of you when you're in a war, so far alliances and vassalization feels like they're mostly for economic and development bonuses with very minor grievance penalties, so they seem to hardly matter war wise aside from making sure that a specific AI doesn't attack you. I also do think that because of the fame victory system, there could be a slightly easier way to design an AI option that makes them play to win the fame victory similar to how Old World has it.