Human in the Loop: Where AI Belongs (and Doesn't) in Social Pt. 2
PEOPLE SHAPING THE CONVERSATION
thank you for your interest
AGENDA
Overview
AI-generated content surpassed human-written content online for the first time in 2025. Agentic AI is now moving from copywriting assistant to actual operator of social accounts, with 58% of enterprise marketing teams already piloting it on at least one channel. And yet, nearly a third of consumers say they're less likely to choose a brand that uses AI ads. The tension isn't resolving — it's intensifying. As AI takes on more social work, the teams that win won't be using it most, but use it most intentionally. And when the scale of AI erodes brand trust, flattens diverse voices, and excludes the people who should be designing these guardrails, leaders must pause to consider what they're willing to sacrifice for efficiency.
Join us for a real conversation about where that line sits, and how senior leaders are deciding what stays human-led, what AI can handle outright, and what lives in the messy middle everyone's still figuring out.
The conversation
- What tasks should AI own outright, what should stay human-led, and where's the messy middle everyone's still figuring out? As senior leaders make this call, you're essentially deciding which decisions get delegated to algorithms and which require judgment, cultural nuance, and accountability. How are you thinking about the tradeoff between speed and control?
- As more brands lean on the same AI tools, content is starting to look and sound the same across the board. When does "efficient" become "generic" and how do you ensure your brand's voice remains distinct?
- Expanding on the above, how do you protect brand trust while still using AI to scale output? Where have you seen AI use backfire or quietly erode engagement? What data are you watching beyond engagement metrics, and how do you catch reputation erosion before it becomes a crisis?
- Who Interprets What AI Detects? AI tools are getting better at spotting signals in real-time. But between detection and action, someone has to decide what a signal means and whether it's worth acting on. How do you keep that interpretation layer human-led without getting bottlenecked? And how do you inform your entire organisation, not just the social team, of these insights quickly?