... Artificial intelligence in serious games uses techniques that are well demonstrated in the entertainmentgamesector-mostnotablyintheformofreflexagents.Thedifference,understandably so,isthatseriousgameproducerscoercetheAIintodirectingrecognizedlearningaidsasopposed tostrivingforsuperiorentertainment. Bellotti,Berta,DeGloria,andPrimavera(2009)proposea gamearchitecturethatincludesAIforeventgeneration,NPCactivation(spawning)andasConnolly, Stansfield,andHainey(2011)purport,appropriatehinting.Thenotionofpersonalizedagent(NPC) responses (Thompson et al., 2010), sometimes directed by an NPC's own agenda (Barab et al., 2012),isalsoanAIpracticeappliedinseriousgames.Generatingusermodelsfromplayeractions, capturedandstoredinadatabase,isanotherconventionalAItechniqueforseriousgames.These usermodelsmanifestthemselvesinthegame'sabilitytoharmonizewithplayerlearningstylesby adapting the game's content presentation (Hwang, Sung, Hung, Huang, & Tsai, 2012;Soflano, Connolly,&Hainey,2015).Bellotti,Berta,DeGloria,D'Ursi,andFiore(2012)useseriousgame adaptivityforadjustingagame'sdifficultylevelaccordingtotheplayer'srecordedability.Thereare otherauthorswho,withoutimplementingAIintheirgames,recognizeandproclaimthepotential ofAI(specificallyusermodelling)toassistwithadaptivefeedback (Cheng,Lin,&She,2015)that willinturnbenefitscaffoldinganddebriefing (Ke,2008).Allinall,theauthorsareconfidentintheir stancethatAIshouldhavemorethanjustatoeholdintheseriousgamesarena.Arecentliterature review(Frutos-Pascual&Zapirain,2017)ontheuseofAIinseriousgamesshowsthatrecentwork inthisarea,aimstotestwhethervariousAItechniquescanbeimposedonseriousgames.Inthe processhowever,manyofthereportedresearchcasesfocusondevelopingAIrichgames,oftenat thecostofequallyimportantfidelity. ...