Aug 9 – Sep 8, 2026 · 5 items
The AI week, for a designer
A real edition, written for: Designer
The one thing
ChatGPT can now operate browsers and spreadsheets directly, which changes what you can hand it during a concept round rather than just what you can ask it.
ChatGPT's new model can drive your browser and spreadsheets, not just chat
OpenAI released GPT-6 Astra, a frontier model built to operate computers directly. It works across browsers, spreadsheets, websites and desktop applications. It can fill forms, update CRM records, run web research and produce documents. OpenAI says this reduces the need for hand-built connectors to each business system. President Greg Brockman told a press briefing that the company is now in the AGI era. Rollout starts Thursday for enterprise customers in the Daybreak gated program. Paid ChatGPT tiers, the OpenAI API, AWS Bedrock and Microsoft Azure follow in coming days. Brockman argued buyers should compare price per completed task rather than per token. OpenAI omitted GDPval, its own benchmark for real-world occupational work. OpenAI also paused some frontier training for about two weeks after the Hugging Face incident and tightened infrastructure controls.
Why it matters for you
You only use ChatGPT, and this changes what you can hand it. Multi-step jobs now run inside apps rather than coming back as text you retype.
- Concept rounds shrink when the model does the tedious steps, not just the writing
- Needs a paid ChatGPT tier, and enterprise access is gated at first
- At 251-2500 people your IT team likely decides which tier lands, so ask before assuming
Try this
Have ChatGPT turn a round of messy campaign feedback into a prioritised change list for your next layout.
Paste this into your AI tool
Here is feedback on a campaign layout from several people: [paste the comments, emails or Slack messages exactly as they arrived]. Group them into: changes I must make, changes that contradict each other, and vague comments I need to ask about. For each contradiction, name the two people and what they each want. For each vague comment, write the one question I should ask to make it actionable. Keep it short enough to read before a call.
Gartner expects most ad money to flow through AI buying platforms by 2028
Gartner expects most advertising money to move through self-serve platforms where AI shapes buying, costs and outcomes by 2028. Eric Schmitt, a Gartner analyst, warns marketers against handing those systems too much control. He argues the platforms' AI serves the seller, not the buyer. Recommendations often amount to advising advertisers to spend more. Schmitt says a human should stay in the loop on budget decisions. He notes the platforms still do not measure or work well across each other. His advice is to cut the number of variables and focus on the largest platforms. He also suggests bringing finance colleagues into the assessment. Independent measurement providers should check campaign results rather than platform reporting. Junior staff can operate the consoles, but experienced oversight remains necessary.
Why it matters for you
You make the campaign visuals that these platforms will chop, test and reassemble. What performs will increasingly be chosen by a machine optimising for the seller.
- Assets built for one hero placement fare badly when a platform auto-generates variants
- The advice is to cut the number of variables, which means fewer, sharper asset systems
- Independent measurement, not platform reporting, is what should judge your creative
Try this
Pick one recent campaign and list which of its assets survive automated resizing and cropping.
Paste this into your AI tool
I design campaign visuals. Here is a description of one recent campaign's asset set: [list the formats, sizes and where each ran]. For each asset, tell me what breaks first if a self-serve ad platform auto-crops or auto-resizes it: logo, headline, focal subject or contrast. Then tell me the three design decisions that would make the whole set safer under automated resizing. Be concrete about placement of type and safe margins, not general principles.
EU AI Act disclosure duties are now enforceable
The European Union's AI Act moved into enforcement on 2 August 2026, with disclosure duties now applying to chatbots and to content produced by AI. The EU's AI Office can request information from covered companies and ask for access to their models, though it has not yet pursued anyone for misconduct. Anthropic, Google, Meta, OpenAI and Microsoft have each described compliance steps, including watermarking generated text. Rules for high-risk uses such as education, biometrics and migration arrive only in December 2027 and August 2028.
Why it matters for you
Swiss agencies and media firms sell into the EU, so campaign work aimed there is covered. AI-generated visuals and copy now carry disclosure obligations.
- Labelling duties sit with whoever puts the content in front of EU users, not just the model vendor
- Vendors are adding watermarking to generated text, which may show up in your delivered files
- High-risk rules land much later, so this is disclosure hygiene rather than a redesign
Try this
Ask whoever signs off your campaign deliverables whether AI-generated visuals need a disclosure line for EU markets.
Sources
Nvidia is buying Hugging Face, the main home of open image and text models
NVIDIA has agreed to acquire Hugging Face, the open model sharing platform. Chief executive Jensen Huang announced the deal on NVIDIA's blog. He put the price at just under $13 billion. NVIDIA says Hugging Face will stay open to the whole AI industry. Developers will keep their choice of models, frameworks, clouds and chips. NVIDIA compute will not be required to build or deploy through the platform. Hugging Face will continue to support open source and open weight models from all builders. NVIDIA says its infrastructure and engineering can improve reliability, safety, model evaluation and deployment. The Hugging Face team and its brand will remain. No closing date or regulatory timeline was given.
Why it matters for you
Most free image generators and design plugins pull their models from Hugging Face. Ownership of that hub now sits with a chip company.
- Nvidia says the platform stays open and its compute stays optional
- The deal needs regulatory approval, so nothing changes for your tools this month
- Worth knowing which of your generative plugins depend on a hub you do not control
Try this
List the generative tools in your design workflow and note which are open models versus paid vendors.
Paste this into your AI tool
I am a brand and digital designer. Here are the AI features and plugins I currently use in my work: [list them, including anything built into design software]. For each, tell me whether it most likely runs on an open downloadable model or a closed paid service, and say how confident you are. Flag any where I should check the terms before using output in client-facing brand work. Keep it to one line per tool.
Sources
Meta scrapped a second layoff round after AI agents underdelivered
Meta cancelled the second wave of layoffs planned under an internal reorganisation that would have shifted much of the daily work of thousands of employees onto AI systems overseen by small expert teams. The company still cut about a tenth of its staff in May, but internal measures showed autonomous AI agents were not delivering the expected productivity, while technical and security incidents rose. Zuckerberg told staff in July that the pace of AI agent progress had been misjudged.
Why it matters for you
You work in media, where the same automation pitch is being made about creative roles. Here is a large firm's own numbers showing it going wrong.
- Output volume rose sharply while incidents and fix times rose faster
- Useful counterweight when someone proposes replacing a review step with an agent
- Zuckerberg's admission that the pace was misjudged is the quotable part
Try this
Note the two review steps in your own process that catch the most errors before handover.
Paste this into your AI tool
I design brand and digital assets: identity, campaign visuals and web layouts. Here is my current process from brief to handover: [describe the steps, including who reviews what]. Identify the two review steps that most likely catch expensive mistakes, and explain what specifically would go wrong without each. Then tell me which single step in my process is the safest to automate and why. Be blunt about the risky ones.
Sources
Build this
Every week, one small thing to build with AI in something you actually care about. No work in it. Five minutes to set up, and worth keeping if it earns a second run.
A finished personal identity kit — one type pairing, one colour set, one grid — built only from the fonts and swatches already on your machine.
Five minutes to set up
I want a small personal visual identity for [what it is for: a portfolio site, a side project, a print zine]. Use only what I already have: [list the typefaces installed on my machine and any colours or swatches I already use]. Rules: no font I do not already own, no paid licence, no suggestion to buy or download anything. Choose one type pairing from my list with specific weights and sizes, one colour set with roles named, and one grid with column count and margins. Explain in one line why each rejected option lost. Output it as a spec I can build from as it stands.
Yours arrives Thursday.
This one was written for a designer. Tell us what you do and the next one is written for you — same news, your job, once a week.