How Organizations Communicate about AI Internally: The New Challenges of Workplace AI

Transcript

1. IT SLOAN SCHOO F MANAGEMENT By Emery Fertic ‘28 Professor Melissa Webster Fall 2025 UROP How Organizations Communicate about AI Internally: The New Challenges of Workplace AI
2. 2 Introductio n Emery Fertic Class of 2028 Mechanical Engineering and AI
3. IT SLOAN SCHOO F MANAGEMENT Acknowledgements : Melissa Webster My research group: Deeksha Gupta and Coco Zhang Ben Fertic 3
4. AGENDA 1. Research Question & Methodology 2. Findings 3. Key takeaways 4. Q&A 4
5. 5 1 Research Question & Methodology
6. How does internal communication about AI differ from communication about previous workplace technologies, and what strategies are organizations using to address those differences? 6
7. 7 Adoption / Enablement Protection / Governance - Why the technology is being introduced - How employees should use it - Data privacy and security - Compliance and approved usage - Risk management What do companies typically communicate during technology adoption? 1
8. Research Methodology Primary Sources: • Attempted outreach across multiple organizations • Conversations were limited due to the sensitivity and evolving nature of internal AI communication practices Secondary Sources: • Internal communication examples • Corporate policy/guidance documents • News reporting • Industry reports and surveys • Academic literature • Conference talks and panels • Generative AI research tools
9. 2 Findings ➔ Communicating about AI presents new challenges compared to other tech ➔ Companies are using four communication strategies
10. 2 Findings ➔ Communicating about AI presents new challenges compared to other tech ➔ Companies are using four communication strategies
11. Three different challenges Employee uncertainty More ongoing communication Explicit guardrails
12. Three different challenges Employee uncertainty More ongoing communication Explicit guardrails
13. How technology was communicated before AI: Focus on: • training • Implementation • process change How technology is communicated now which leads to distrust: Has to address the following: • job displacement fears • acceptable use confusion • privacy concerns • surveillance concerns • changing skill expectations Companies face employee uncertainties , respond by emphasizing trust Organizations respond through trust-building Organizations increasingly try to communicate: • transparency • responsible AI • ethical AI • human oversight • validation requirements
14. 14 AI shapes how employees interpret trust, security, and even their role. Amazon (2025–2026) Google (2026) Meta (Apr–May 2026) 1 Sources: Reuters (2026); Business Insider (2026); Fast Company (2026); HR Executive (2026) Employees reacted negatively to AI-training data collection and tracking, a case where communication created distrust rather than reassurance. Employees reported fear of replacement and pressure to use AI tools despite quality concerns. Performance review integration around AI adoption shows how AI becomes tied to employee identity/performance.
15. Employees described Meta as an “Employee Data Extraction Factory.” Employees worried the company was “using them to train their eventual bot replacements.” – Employees when asked about Meta’s new policy that will track their computers and record data on how they do their work 1 Source: Reuters, ‘Meta employees organize protest against mouse-tracking tech,’ May 2026.
16. Three different challenges Employee uncertainty More ongoing communication Explicit guardrails
17. - One-time rollout announcements - Standardized training sessions - IT support documentation - Top-down deployment campaigns More ongoing communication 17 - Ongoing feedback loops - Internal AI communities - Manager modeling - Continuous experimentation & updates *AI communication is increasingly conversational and iterative rather than static and deployment-focused. How technology is communicated now: How technology was communicated before AI: 2
18. 18 AI communication is becoming much more ongoing and interactive Google - AI-moderated all-hands meetings show experimentation with new communication formats themselves. Microsoft Manager Modeling - Microsoft found that employees reported higher trust in AI when managers actively modeled AI use, highlighting the importance of ongoing communication and reinforcement. 2 Sources: Microsoft Work Trend Index (2026); Google internal AI communication reporting/disclosures (2026)
19. Three different challenges Employee uncertainty More ongoing communication Explicit guardrails
20. Explicit Guardrails 20 Why AI requires more guardrails: Employees need: - behavioral guidance - disclosure expectations - privacy rules - approved-use policies, - human oversight requirements - Unlike many previous workplace technologies, generative AI reached consumers before organizations established formal workplace policies. What organizations are doing now: Organizations are: - limiting usage - retroactively governing behavior - responding to existing employee adoption Instead of: - selecting enterprise tools first - rolling them out top-down 3
21. 21 Companies are actively defining boundaries, responsibilities, and acceptable behavior JPMorgan - JPMorgan introduced strict approved-use policies and governance controls around employee AI usage, emphasizing oversight, compliance, and data protection. IBM - IBM created an internal “AI license to drive” certification program to train employees on responsible AI use, privacy, security, and governance before building AI systems. Microsoft - Microsoft framed its Copilot rollout around sensitivity labels, data loss prevention (DLP), and responsible AI guidance to ensure secure internal AI use. 3 Sources: IBM, JPMorgan, and Microsoft AI governance documentation/reporting (2024–2026).
22. Organizational culture and manager support contribute more than mindset alone to AI impact Contribution to AI impact (%) (2026) Source: McKinsey Global Survey on AI, Superagency in the Workplace (2026).
23. 2 Findings ➔ Communicating about AI presents new challenges compared to other tech ➔ Companies are using four communication strategies
24. Four communication strategies Enablement & Training Governance & Trust Employee participation Mandates/surveillance
25. Four communication strategies Enablement & Training Governance & Trust Employee participation Mandates & surveillance 1
26. 58% say they’re producing work they couldn’t have a year ago. 66% of AI users say AI has allowed them to spend more time on high-value work. Enablement & training Companies communicate AI as a skill employees must learn, not just a tool rollout. 1
27. 27 Used Viva for awareness, training, feedback, employee champions, and usage tracking across 300,000+ employees/vendors. Microsoft Rolling AI tools/training to its U.S. associate base, including translation and task tools that cut shift planning from 90 to 30 minutes. Walmart Enablement & training examples 1
28. Four communication strategies Enablement & Training Governance & Trust Employee participation Mandates & surveillance 2
29. Only 26% of AI users say leadership is clearly and consistently aligned on AI. Governance & trust Governance & trust Companies emphasize rules, data protection, and “safe AI.” 2
30. 30 Frames Copilot rollout around sensitivity labels, DLP, works councils, and employee awareness. Microsoft Uses an internal “AI license to drive” certification, so employees understand privacy, security, and enterprise integration before building agents. IBM Governance & trust examples 2
31. Four communication strategies Enablement & Training Governance & Trust Employee participation Mandates & surveillance 3
32. 22 point lift in critical thinking 17 point lift in AI value Employee Participation The best cases treat employees as co-designers, not just recipients. 30 point lift in trust in AI 3
33. 33 Used listening sessions, satisfaction surveys, Viva Engage, employee champions, and Viva Pulse feedback to refine adoption. Microsoft Launched internal AI tools, reached 80% internal adoption, and deployed 750+ GPTs across legal, research, manufacturing, and commercial functions. Moderna Employee participation examples 3
34. Four communication strategies Enablement & Training Governance & Trust Employee participation Mandates & surveillance
35. Only 13% say they are rewarded for reinvention with AI even if results aren’t met 65% of AI users fear falling behind if they don’t adapt with AI quickly Mandates & surveillance Some companies communicate AI through pressure, monitoring, or usage targets, which can hurt trust. 4
36. 36 Employees reported pressure to use AI and concern about tracking/low-quality outputs; recent reporting also says Amazon tracks usage, token consumption, and weekly adoption goals. Amazon Reportedly began capturing employee mouse/keyboard/screen activity for AI training, causing privacy backlash. Meta Mandates & surveillance examples 4
37. 37 The big 5 tech companies emphasize different communication types Enablement & training Governance & trust Employee Participation Mandates/ Surveillance Google ✓ ✓ ✓ Microsoft ✓ ✓ ✓ Apple ✓ Amazon ✓ Meta ✓
38. 38 3 Key Takeaways
39. 3 Key Takeaways from spring 2026 research 39 Organizations are balancing enablement with governance. Internal communication strategy increasingly shapes AI adoption outcomes. AI communication is no longer just about teaching employees how to use a tool. 1. 2. 3.
40. Future research directions Accessing primary organizational perspectives: 1. How are organizations internally discussing employee concerns around AI, including job displacement, surveillance, and trust? 2. What communication strategies do leaders believe are most effective for increasing AI adoption while maintaining employee trust? Measuring long-term employee impact: 1. How does AI communication affect employee trust, morale, and perceptions of job security over time? 2. Do different communication strategies influence adoption rates, burnout, or organizational trust differently? Comparing industries and organizational cultures 1. How do AI communication strategies differ across industries such as finance, healthcare, manufacturing, and technology? 2. Are highly regulated industries more governance-focused while tech firms emphasize experimentation and adoption? 40
41. IT SLOAN SCHOO F MANAGEMENT Special thank you to Melissa Webster! 41
42. 42 4 Q/A
43. 43 5 Appendix
44. Four different challenges Employee uncertainty More ongoing communication Explicit guardrails More emphasis on trust
45. - job displacement fears - concerns about monitoring - uncertainty around AI expectations - skepticism toward leadership intentions “ Employee uncertainty causes for a bigger emphasis on trust 45 Employee uncertainty creates distrust Organizations respond through trust-building Organizations increasingly try to communicate: - transparency - responsible AI - ethical AI - human oversight - validation requirements 1
46. Employee participation 46 The best cases treat employees as co-designers, not just recipients. When managers actively modeled AI use, employees reported: 17 point lift in AI value 22 point lift in critical thinking 30 point lift in trust in AI 3
47. Governance & trust 47 Companies emphasize rules, data protection, and “safe AI. Only 26% of AI users say leadership is clearly and consistently aligned on AI. 2
48. Mandates & surveillance 48 Some companies communicate AI through pressure, monitoring, or usage targets, which can hurt trust. 65% of AI users fear falling behind if they don’t adapt with AI quickly Only 13% say they are rewarded for reinvention with AI even if results aren’t met 4