AI Budgets Are Climbing While Trust in AI Ads Is Falling: This Week in Social Media for Agencies

The week in brief:

• A global Auros/Talker Research survey of 4,000 adults in the US, UK and Australia found 53% of Americans (58% in the UK, 59% in Australia) do not trust products advertised with AI to be described accurately, and 44% say they go out of their way to buy from brands that skip AI in their marketing.
• Instagram's "AI-generated profile" label (renamed from "AI creator" on August 31) is moving toward active enforcement, cutting Reels and Explore reach for accounts with an undisclosed AI persona.
• Meta dropped the rule requiring Threads advertisers to hold a matching Instagram username, so a brand with no Instagram presence can now run Threads ads directly from its own profile.
• A Huble Digital survey of 393 HubSpot customers at HubSpot's UNBOUND conference (September 16 to 18) found 70% plan to increase AI spend over the next 12 months, with 23% planning to spend "much more."

Put these four stories side by side and they point in opposite directions on purpose. Marketers are putting more budget behind AI than ever, and the platforms are putting more rules around disclosing it, while the audience on the other end is telling pollsters, in plain language, that they don't trust it. None of that is a reason to pull back on AI. It's a reason to get specific about where it's doing the work and where a person needs to still be the one signing off.

Consumers don't trust AI-written ads, but they like using AI themselves

Auros, an AI-commerce company, commissioned Talker Research to poll 4,000 adults across the US, UK and Australia on how they feel about AI showing up in retail and advertising (eMarketer, MediaPost). The headline number: a majority in every country surveyed said they don't trust a product advertised using AI to be described accurately, 53% of Americans, 58% of Brits, 59% of Australians. Dig one layer deeper and it gets more pointed. 36% said a product pushed with AI-generated ad content reads as lower quality to them, and 44% said they actively go out of their way to shop with brands that aren't using AI in their marketing at all.

Here's the part agencies should actually sit with: the same respondents weren't anti-AI across the board. A separate finding from the same study said AI has genuinely helped people shop smarter when they're the one steering it, researching options, comparing prices, narrowing a decision. The distrust shows up specifically when AI is doing the persuading on the brand's side, not when a shopper is using AI as their own tool. That's a useful, narrow read: the backlash isn't "AI touched this," it's "nobody stopped to check this before it reached me."

For client work, that's an argument you can make in plain terms at the next strategy call. Drafting a caption with AI and having a person edit, fact-check and approve it before it posts is a completely different trust proposition than letting a model's first draft go straight out the door. The gap the survey is measuring isn't a technology problem. It's a supervision problem, and supervision is exactly what a retainer is supposed to buy.

Instagram starts treating an undisclosed AI persona as a reach problem

Instagram renamed its "AI creator" label to "AI-generated profile" on August 31 and said accounts built around an AI-generated person that skip the label will stop being recommended to non-followers in Reels and Explore once detection rolls out account by account (TechCrunch). Accounts that proactively add the label see no change to their reach. The scope is narrower than it sounds: this is about profiles built around an AI persona, not about creators who use AI tools for editing, captions or graphics, so a client's normal AI-assisted workflow isn't what this targets.

If any client account runs an AI spokesperson, mascot, or fully synthetic "creator" persona, the label needs to go on now rather than after a reach drop shows up in a monthly report with no obvious cause. This sits next to, but is separate from, the state and federal disclosure laws already in effect for AI-generated performers in paid ads; our guide to those rules covers the California and New York statutes and the FTC's Endorsement Guides, which are legal disclosure requirements, not a platform ranking policy. Instagram's label is Meta's own reach lever, and it's worth checking separately from legal compliance, because a label that satisfies the FTC doesn't automatically protect an account's Explore reach.

Threads ads no longer require an Instagram account at all

Meta also dropped a requirement that used to trip up a specific kind of client: advertisers needed a Threads username that matched their Instagram handle to run ads on Threads. That's gone. Starting with a rollout that began in September and is still reaching accounts gradually, a business with no Instagram presence at all can now advertise directly from its own Threads profile (Social Media Today). Existing ad setups in Meta Business Suite keep running as-is, including if a Threads or Instagram username changes later.

This matters for a narrow but real client type: the brand that wants a presence on Threads specifically (community, commentary, a text-first audience) without first building out an Instagram account it doesn't actually need. Before this, that client was stuck either standing up a token Instagram profile just to unlock ads or skipping Threads ads entirely. Check the client's ad account in Meta Business Suite before promising a launch date, since the rollout is gradual and not every account has it yet.

HubSpot customers are planning to spend more on AI, not less

Huble Digital, a HubSpot Elite Solutions Partner, surveyed 393 HubSpot customers at HubSpot's UNBOUND conference in Boston (September 16 to 18) and found 70% plan to increase their AI spend over the next 12 months, with 23% planning to spend "much more" (Complete AI Training, reported in agilebrandguide.com's October 17 roundup). Intent to spend more or much more ran lower among US customers (68%) than the rest of the sample (77%). Adoption is already high: 66% use HubSpot's built-in Breeze AI assistant, and 61% use two or more AI tools in their business daily, with Claude, ChatGPT and Copilot named most often as the outside tools alongside HubSpot's own. Huble's own caveat is worth repeating: its CEO argues the results depend on whether a company's CRM data is clean enough to make the AI useful in the first place, not just on how much gets spent.

Read against the Auros survey above, this is the same week telling two sides of one story. The money is moving toward AI faster than trust in it is. The agencies that come out ahead aren't the ones spending the most on AI tools, they're the ones that can show a client exactly where a person checked the work before it reached the client's audience, the same distinction a good monthly report already has to make between what the tool did and what the account actually earned.

What to actually do with this before next week

Audit any client account running an AI-generated persona and get Instagram's label on it this week, not after a reach report looks strange. Separately, if that same account runs paid ads with an AI-generated voice or performer, check it against the California and New York disclosure rules, since those are a different requirement with real penalties attached. If a client wants Threads ads and doesn't already run Instagram, this is the moment to raise it, since the blocker that used to stop that conversation is gone. And in any pitch or renewal conversation that touches AI, lead with where a person is checking the AI's work, not with how much AI the account uses. That's the half of this week's news that's actually a selling point.

FAQ

Do consumers trust AI-generated ads?

No, not by default. A 2026 Auros/Talker Research survey of 4,000 adults across the US, UK and Australia found a majority in every country distrust AI-advertised products to be described accurately, though the same respondents said AI tools help them when they are the ones using it to shop.

Does Instagram's AI-generated profile label apply to agencies that just use AI to write captions or edit photos?

No. Instagram's label and reach penalty target profiles built around a synthetic AI persona, a fully AI-generated "creator" or spokesperson. Using AI tools for captions, photo edits or graphics on a real person's account does not trigger it.

Can a brand run Threads ads without having an Instagram account?

Yes, as of Meta's September 2026 rollout, a business with no Instagram profile can advertise directly from its own Threads account. The change is rolling out gradually, so confirm it is live on a specific client's ad account before planning a launch date.

Is Instagram's AI label the same as the AI disclosure laws in California and New York?

No. Instagram's label is a platform ranking policy that affects Reels and Explore reach. California and New York's synthetic performer disclosure laws, and the FTC's Endorsement Guides, are separate legal requirements with their own penalties for undisclosed AI performers in ads.

How much are marketers planning to increase AI spending?

A Huble Digital survey of 393 HubSpot customers found 70% plan to increase AI spend over the next 12 months, including 23% planning to spend "much more," based on a poll taken at HubSpot's UNBOUND conference in September 2026.