When a tool makes production faster, the temptation is to count the extra output. Canva co-founder and chief product officer Cameron Adams used his August recap of AI Vision Sydney to ask a different question: what should people keep responsibility for when software can do more of the making?
In the recap published on August 24, Adams described an event bringing together creative and business practitioners to discuss AI and human work. The themes included building better systems around AI, protecting brand consistency, tracing creative provenance and retaining human judgment. This article draws on that public account; it is not a GFT interview or firsthand event report.
The founder’s framing is especially relevant to a design platform. Faster production can make a tool more useful, but customers still need a reason to prefer the result. A campaign that is technically competent yet indistinguishable from its competitors may not solve the customer’s actual communication problem.
GFT analysis: the practical distinction is between making an option and choosing an option. A business can automate parts of the first task while remaining accountable for the second. Someone must still decide whether a design represents the organisation, whether a claim is supported and whether the finished work helps its intended audience.
For a creative agency or small brand, one useful experiment is to measure revision quality alongside production speed. Ask reviewers to identify what improved, what became generic and where factual or brand checks were needed. Time saved in a first draft can disappear if the final review becomes longer and less reliable.
A clear record of sources, permissions and editorial decisions can also become part of the product’s credibility. That is particularly important for publications using real people and real events: an attractive visual is not a substitute for an authentic photograph or an accurate caption. Adams’s August reflection provides a company founder’s perspective on the issue. The broader business lesson is a question worth testing: does automation leave a team with more room for good judgment, or merely more material to judge?