The advent of large language models is fundamentally reshaping the online search landscape, challenging Google's two-decade dominance and opening a new competitive phase between traditional search engines and generative artificial intelligence platforms. This thesis examines the phenomenon of the "Search War" in the era of artificial intelligence, with particular attention to its impact on user behaviour and the digital commerce ecosystem. The aim of this research is to investigate whether, and to what extent, AI-powered search engines are replacing Google as the primary entry point for commercial search, and to assess the implications of this shift for digital marketing business models, organic traffic, and user trust. The research adopts a mixed-methods approach. The first, theoretical phase reconstructs the historical and technological evolution of search engines through an analysis of the academic literature and other relevant sources, from the web directories of the 1990s to the current generative AI models. It also identifies the main actors in the new Search War, namely Google Gemini, ChatGPT, and Perplexity AI. The second, empirical phase consists of comparative tests conducted using Google Search, ChatGPT, and Arc Browser through two real-world use cases: a transactional search for sports footwear and an advisory search for skincare products. The performance of each platform is evaluated through a matrix of key performance indicators (KPIs). The study is complemented by a qualitative focus group conducted at the University of Pavia with eleven students enrolled in the Digital Communication course. The research results demonstrate that AI-powered search engines achieve better performance than traditional search engines in terms of relevance, absence of advertising noise, and time efficiency. They also indicate that users already adopt these tools as their primary entry point for complex commercial searches. The focus group revealed that this adoption is driven by such a high level of perceived convenience that it substantially reduces critical reflection. Furthermore, awareness of the negative effects of the zero-click phenomenon does not influence users' behaviour. This thesis concludes that the ongoing transformation is not only technological but also cognitive and behavioural. The challenges associated with algorithmic opacity and its impact on the digital content ecosystem require structural measures aimed at improving regulation and transparency, together with new visibility strategies based on Generative Engine Optimization (GEO).
L’avvento dei modelli linguistici di grandi dimensioni sta ridefinendo in modo profondo il panorama della ricerca online, mettendo in discussione il primato ventennale di Google e aprendo una nuova fase competitiva tra motori di ricerca tradizionali e piattaforme di intelligenza artificiale generativa. Questo elaborato analizza il fenomeno della “Search War” nell’era dell’intelligenza artificiale, con particolare attenzione alle sue ricadute sul comportamento degli utenti e sull’ecosistema del commercio digitale. L’obiettivo della ricerca è verificare se e in che misura i motori di ricerca basati sull’IA stiano sostituendo Google come punto di accesso primario alla ricerca commerciale, e quali conseguenze questo cambiamento produca sui modelli di business del marketing digitale, sul traffico organico e sulla fiducia degli utenti. La ricerca adotta un approccio misto. Una prima fase teorica ricostruisce l’evoluzione storica e tecnologica dei motori di ricerca attraverso l’analisi della letteratura accademica e delle fonti bibliografiche disponibili, dai directory degli anni Novanta fino ai modelli generativi attuali, e mappa i principali attori della nuova Search War: Google Gemini, ChatGPT e Perplexity AI. Una seconda fase empirica prevede test comparativi condotti su Google Search, ChatGPT e Arc Browser attraverso due casi d’uso reali: una ricerca transazionale su calzature sportive e una consulenziale su prodotti skincare, valutati tramite una matrice di KPI. La ricerca si completa con un focus group qualitativo svolto presso l’Università di Pavia, con undici studenti di Comunicazione Digitale. I risultati mostrano che i motori AI offrono prestazioni superiori in termini di pertinenza, assenza di rumore pubblicitario ed efficienza temporale, e che gli utenti li adottano già come primo punto d’accesso per ricerche commerciali complesse. Il focus group ha rivelato che questa adozione è guidata da una comodità percepita così elevata da azzerare la riflessione critica, e che la consapevolezza degli effetti negativi del fenomeno zero-clic non modifica il comportamento. Questo elaborato conclude che il cambiamento in atto non è solo tecnologico ma cognitivo e comportamentale, e che le soluzioni all’opacità algoritmica e all’impatto sull’ecosistema dei contenuti richiedono interventi strutturali sul piano della regolamentazione e della trasparenza, oltre che nuove strategie di visibilità basate sulla GEO (Generative Engine Optimization).
LA “SEARCH WAR” NELL’ERA DELL’INTELLIGENZA ARTIFICIALE
SCOLA, GIADA
2025/2026
Abstract
The advent of large language models is fundamentally reshaping the online search landscape, challenging Google's two-decade dominance and opening a new competitive phase between traditional search engines and generative artificial intelligence platforms. This thesis examines the phenomenon of the "Search War" in the era of artificial intelligence, with particular attention to its impact on user behaviour and the digital commerce ecosystem. The aim of this research is to investigate whether, and to what extent, AI-powered search engines are replacing Google as the primary entry point for commercial search, and to assess the implications of this shift for digital marketing business models, organic traffic, and user trust. The research adopts a mixed-methods approach. The first, theoretical phase reconstructs the historical and technological evolution of search engines through an analysis of the academic literature and other relevant sources, from the web directories of the 1990s to the current generative AI models. It also identifies the main actors in the new Search War, namely Google Gemini, ChatGPT, and Perplexity AI. The second, empirical phase consists of comparative tests conducted using Google Search, ChatGPT, and Arc Browser through two real-world use cases: a transactional search for sports footwear and an advisory search for skincare products. The performance of each platform is evaluated through a matrix of key performance indicators (KPIs). The study is complemented by a qualitative focus group conducted at the University of Pavia with eleven students enrolled in the Digital Communication course. The research results demonstrate that AI-powered search engines achieve better performance than traditional search engines in terms of relevance, absence of advertising noise, and time efficiency. They also indicate that users already adopt these tools as their primary entry point for complex commercial searches. The focus group revealed that this adoption is driven by such a high level of perceived convenience that it substantially reduces critical reflection. Furthermore, awareness of the negative effects of the zero-click phenomenon does not influence users' behaviour. This thesis concludes that the ongoing transformation is not only technological but also cognitive and behavioural. The challenges associated with algorithmic opacity and its impact on the digital content ecosystem require structural measures aimed at improving regulation and transparency, together with new visibility strategies based on Generative Engine Optimization (GEO).| File | Dimensione | Formato | |
|---|---|---|---|
|
SCOLA GIADA TESI.pdf
embargo fino al 25/04/2027
Dimensione
5.18 MB
Formato
Adobe PDF
|
5.18 MB | Adobe PDF | Richiedi una copia |
È consentito all'utente scaricare e condividere i documenti disponibili a testo pieno in UNITESI UNIPV nel rispetto della licenza Creative Commons del tipo CC BY NC ND.
Per maggiori informazioni e per verifiche sull'eventuale disponibilità del file scrivere a: [email protected].
https://hdl.handle.net/20.500.14239/37182