Be careful who you listen to on AI, and a push for no-manipulation clauses in speaker contracts

A prominent social media industry figure set off a wide discussion this week with a simple warning about the AI debate: be careful who you listen to. Everyone in the conversation has hidden motivations and unrecognized biases that shape what they say, what they hope to hear, and what they hold to be true.

Silicon Valley investors, the argument goes, want the public to believe AI agents are on track to be quicker, more capable and less expensive than people, and billionaire backers pouring enormous sums into AI need the bet to pay off. The measured counterview: there is no need to panic, and most jobs are secure for a while longer, but that is no excuse to sit idle while the technology reshapes marketing and media.

The same community is pushing a concrete ethics measure: a no-AI-manipulation clause for speaker contracts. The idea gained traction after two incidents. In one, an event host promoting a webinar used a completely fake AI-generated image of the booked speaker instead of the supplied headshot; the speaker demanded it come down, refused to have her likeness tampered with, and both sides ultimately cancelled the session. In another, a deep-fake AI-generated clip was spliced into a personal interview to make the interviewee appear emotional and distressed, dismissed by the host as just B-roll.

The response from fellow speakers and event hosts was strong enough that a recommended contract clause, drafted with input from a lawyer, has been published as a freely shareable document. The underlying principle, as one affected marketer put it in a widely shared post, is that ethical AI use is not optional: altering someone's image or emotional state without permission crosses a line, and making light of it afterward makes it worse.

In platform news, Instagram is publicly testing a feature that lets users see and control their own algorithm. A small percentage of US users can tune recommendations by adding and removing interest topics, starting with Reels and expanding to Explore, with a version also being explored for Threads. Threads, meanwhile, has introduced ghost posts, an ephemeral format Meta says will encourage unfiltered thoughts and experimental content, though skeptics note that more personal sharing also means more data for ad targeting.

Also announced: a dedicated AI business conference will run in April 2026 in Anaheim as a sub-conference of a long-running social media marketing event, with standalone and virtual tickets, and a free live training on social media strategy for 2026 is set for November 5, promising practical tactics over keeping up with mega creators.