Impact of Artificial Intelligence and Social Media on Interpersonal Communication among University Students

María Fernanda Aveiga Loor · Juan Carlos Solano Jiménez (Presenter) · Valeria del Rosario Herrera Salazar · Juan José Arciniega Cuenca

Universidad Nacional de Loja, Facultad de la Energía, las Industrias y los Recursos Naturales No Renovables, Loja 110111, Ecuador

📊 N = 108 University Students (18–35 yrs) 🛡️ Cronbach's α = 0.84 · R v4.3

90.7%

Multi-Daily Access

98/108 connect multiple times/day

54.0%

> 3 Hours Daily

Intensive social media immersion

88.9%

AI Tool Adoption

96/108 actively use generative AI

52.0%

Perceived Insincerity

Rate AI messages as less authentic

84.0%

Face-to-Face Preference

For sensitive & emotional topics

Daily Connection Duration

n = 108

Continuous immersion: Over half of respondents (54.0%) exceed 3 hours daily on social platforms, and 36.7% spend 1–3 hours.

AI Tools & Platform Dominance

Multi-Response

Ecosystem hierarchy: WhatsApp reaches 100% saturation; ChatGPT (60.2%) and Gemini (39.8%) lead generative AI adoption.

Functional Boundary of AI

Academic vs. Personal

Selective boundary: Students embrace AI for academic text editing (68.5%) and summarizing (61.1%), but restrict it in personal messaging (14.8%).

Perceived Emotional & Cognitive Outcomes

Fig. 3 Data

The Dual Effect: While 41.0% gain efficiency and 25.0% report improved peer contact, nearly 29% experience mental fatigue (19.4%) or immediate-response anxiety (9.3%).

The Efficiency–Authenticity Paradox

Media Richness Theory

Prefer Face-to-Face for Sensitive / Emotional Topics 84.0% (91/108)

Perceive AI-Assisted Messages as Less Authentic / Sincere 52.0% (56/108)

Report Reduced Frequency of In-Person Meetings 46.0% (50/108)

Highlight Functional Speed & Productivity Gains 41.0% (44/108)

Use AI to Draft Personal / Relational Messages 14.8% (16/108)

Theoretical Alignment (Walther & Daft/Lengel): Algorithmic polishing strips spontaneous emotional cues. When ambiguity is high, students overwhelmingly return to face-to-face dialogue (84.0%).

Impact on Authenticity & Physical Encounters

Fig. 4 Data

Key Takeaways for Higher Education Institutions

UNL · Policy & Pedagogy

1. Emotional Digital Literacy: Move beyond technical prompt engineering; train students to recognize when algorithmic mediation erodes relational trust and empathy.

2. Preserving Co-Present University Spaces: With 46.0% reporting fewer physical meetings, curricula should intentionally design collaborative, face-to-face problem-solving workshops.

3. Digital Well-Being & Disconnection Norms: Address hyper-connectivity fatigue (19.4%) and immediate-reply anxiety (9.3%) by establishing healthy asynchronous communication norms.

Interactive 2×2 Contingency Lab (Pearson's χ²)

df = 1 · α = 0.05

Test 1: Daily Connectivity × AI Insincerity

χ² = 7.14 · p = 0.0075 · Cramér's V = 0.26

Test 2: Messaging Reliance × Fewer Visits

χ² = 9.82 · p = 0.0017 · Cramér's V = 0.30

7.14

Pearson χ² (1, N=108)

0.0075

p-value (< 0.01)

0.26

Cramér's V (Effect Size)

Significant

Statistical Decision

Subgroup Comparison (%)

Empirical Association

Interpretation (Test 1): Students connected >3 hours/day are significantly more likely to perceive AI-assisted messages as impersonal or insincere (63.8% vs. 38.0% among moderate users, p = 0.0075). Greater digital exposure sharpens sensitivity to synthetic phrasing.