Umetna inteligenca: primerjalna analiza Copilot, ChatGPT in DeepSeek

Belšak, Kevin

| | SLO | ENG | Cookies and privacy

Bigger font | Smaller font

First page > Show document

Show document

Title:	Umetna inteligenca: primerjalna analiza Copilot, ChatGPT in DeepSeek
Authors:	ID Belšak, Kevin (Author) ID Perko, Igor (Mentor) More about this mentor...
Files:	MAG_Belsak_Kevin_2026.pdf (5,71 MB) MD5: 5B87FF5046F59B11196FB85616A01682
Language:	Slovenian
Work type:	Master's thesis/paper
Typology:	2.09 - Master's Thesis
Organization:	EPF - Faculty of Business and Economics
Abstract:	V magistrskem delu sem primerjal tri priljubljene modele umetne inteligence Microsoft Copilot, ChatGPT in DeepSeek. Zanimalo me je, kako dobro razumejo naravni jezik, kako varno ravnajo s podatki in kako se povezujejo z orodji, ki jih uporabljamo pri vsakdanjem delu. Čeprav so ti modeli vse bolj prisotni na različnih področjih, se uporabniki pogosto sprašujejo, kateri je zares najboljši in ali mu lahko zaupajo svoje podatke. V raziskavi je sodelovalo 65 ljudi, ki so izpolnili spletno anketo. Izkazalo se je, da med modeli obstajajo pomembne razlike. ChatGPT je bil najvišje ocenjen pri razumevanju uporabnikovega namena in ohranjanju konteksta pogovora, uporabniki so z njim tudi najbolj zadovoljni. DeepSeek se je odrezal dobro pri natančnosti odgovorov, kar potrjuje njegovo usmerjenost v tehnične naloge. Microsoft Copilot je pri razumevanju jezika pričakovano zaostal, saj je njegova prava moč drugje v povezanosti z Microsoftovimi orodji. Pri varnosti in zasebnosti uporabniki najbolj zaupajo ChatGPT-ju, vendar kljub temu več kot polovica vanj ne vnaša občutljivih podatkov. DeepSeek je na tem področju najslabše ocenjen, verjetno zaradi pomislekov glede njegovega izvora in preglednosti. Skoraj tretjina vseh sodelujočih ne zaupa nobenemu modelu, kar kaže na splošno zadržanost do umetne inteligence. Pri povezovanju z orodji ima Microsoft Copilot največjo prednost deluje neposredno znotraj Worda, Excela in Teamsov, zato uporabniku ni treba preklapljati med okni. ChatGPT je izjemno prilagodljiv, vendar pogosto zahteva ročno kopiranje in lepljenje. DeepSeek integracije skoraj ne podpira in je namenjen predvsem tehničnim nalogam. Nobeden od modelov ni najboljši v vsem. ChatGPT je odlična vsestranska izbira, Copilot je nepogrešljiv znotraj Microsoftovega okolja, DeepSeek pa pride prav pri zahtevnejših tehničnih opravilih. Izbira je torej odvisna od tega, kaj uporabnik dejansko potrebuje in kakšno okolje uporablja.
Keywords:	Umetna inteligenca, ChatGPT, Microsoft Copilot, DeepSeek, primerjalna analiza, integracija orodij.
Place of publishing:	Maribor
Publisher:	K. Belšak]
Year of publishing:	2026
PID:	20.500.12556/DKUM-97515
UDC:	004.8
COBISS.SI-ID:	280421891
Publication date in DKUM:	04.06.2026
Views:	216
Downloads:	19
Metadata:
Categories:	EPF
:	Copy citation

Average score:	(0 votes)
Your score:	Voting is allowed only for logged in users.
Share:

Hover the mouse pointer over a document title to show the abstract or click on the title to get all document metadata.

Licences

License:	CC BY-SA 4.0, Creative Commons Attribution-ShareAlike 4.0 International

Link:	http://creativecommons.org/licenses/by-sa/4.0/
Description:	This Creative Commons license is very similar to the regular Attribution license, but requires the release of all derivative works under this same license.
Licensing start date:	13.03.2026

Secondary language

Language:	English
Title:	Artificial intelligence: comparative analysis of copilot, chatgpt and deepseek
Abstract:	In my master's thesis, I compared three popular artificial intelligence models Microsoft Copilot, ChatGPT, and DeepSeek. I wanted to find out how well they understand natural language, how safely they handle data, and how they connect with the tools we use in everyday work. Although these models are becoming increasingly common across different fields, users often wonder which one is truly the best and whether they can trust them with their data. 65 people participated in the research by completing an online survey. The results showed significant differences between the models. ChatGPT received the highest ratings for understanding user intent and maintaining conversation context, and users were most satisfied with it overall. DeepSeek performed well in answer accuracy, confirming its focus on technical tasks. Microsoft Copilot lagged behind in language understanding as expected, since its real strength lies elsewhere in its integration with Microsoft tools. When it comes to security and privacy, users trust ChatGPT the most, yet more than half of them avoid using it for sensitive data. DeepSeek was rated lowest in this area, likely due to concerns about its origin and transparency. Nearly a third of all participants don't trust any of the models, which points to a general hesitation towards artificial intelligence. In terms of tool integration, Microsoft Copilot has the biggest advantage it works directly within Word, Excel, and Teams, so users don't have to switch between windows. ChatGPT is extremely flexible but often requires manual copying and pasting. DeepSeek barely supports any integration and is mainly intended for technical tasks. None of the models is best at everything. ChatGPT is an excellent all-round choice, Copilot is indispensable within the Microsoft ecosystem, and DeepSeek comes in handy for more demanding technical tasks. The choice ultimately depends on what the user actually needs and what environment they work in.
Keywords:	Artificial intelligence, ChatGPT, Microsoft Copilot, DeepSeek, comparative analysis, tool integration.

Comments

Leave comment

You must log in to leave a comment.

Comments (0)

0 - 0 / 0

There are no comments!

Back