This week the artificial intelligence marketing agenda was set by transparency and corporate adoption: as Google starts labelling AI use in advertising, OpenAI's new GPT-5.6 family has become the default engine in Microsoft 365 Copilot. Meta, meanwhile, pulled a feature that used users' photographs without permission, giving an important lesson in brand trust. Here are the week's highlights with DijitalPi's commentary from the field.
Google Is Starting to Label Ads Produced With Artificial Intelligence
Google is bringing in a new transparency requirement for image and video ads created with generative artificial intelligence tools. Until now compulsory only for election ads, that labelling will now apply to all advertisers. Advertisers will have to state clearly to the user which content was produced or altered with AI.
DijitalPi's Take
Transparency in advertising is becoming a rule rather than a choice. For e-commerce and retail brands in Türkiye, this change means accounts using AI in product images will be labelled in the coming period; that creates both a risk and an opportunity for consumer trust, because for a correctly positioned brand transparency can in fact turn into a trust signal. Our recommendation: document with your creative teams now which images were produced with AI, and when labelling arrives track your campaign performance by segment.
GPT-5.6 Has Become the Default Model in Microsoft 365 Copilot
OpenAI's new model family GPT-5.6 has risen to the position of preferred model in Microsoft's corporate productivity suite Copilot 365. Promising stronger performance in tools such as Word, Excel, PowerPoint and Chat, this integration is continuing despite debate about the partnership between the two companies.
DijitalPi's Take
Corporate artificial intelligence is no longer a laboratory experiment but part of the daily workflow. For B2B marketing and sales teams, this means the quality of proposals, reports and content prepared through Copilot jumping overnight; the difference will be felt immediately in the agency and consultancy sector in particular, where content production speed turns into a competitive advantage. Our recommendation: review your team's Copilot 365 usage habits and fold the capabilities GPT-5.6 brings (long document analysis in particular) into your existing workflows now.
Meta Has Shut Down the Muse Image Feature That Used Photographs Without Permission
Following the reaction it received, Meta has removed the Muse Image feature, which allowed photographs on public Instagram accounts to be used in AI image production through tagging without permission being obtained. The company stated the feature's aim was to offer a creative tool but that it did not meet expectations.
DijitalPi's Take
Every shortcut that plays with user data eventually bills brand reputation. For brands doing social media and influencer marketing, this incident shows they must read platform-based AI tools' data usage policies line by line before folding them into their campaigns; for brands building a UGC (user-generated content) strategy in particular, the permission mechanism should now be part of the marketing plan. Our recommendation: get approval from your legal and data privacy team before using third-party AI tools in your campaigns, and keep an 'AI-driven backlash' scenario in your crisis communication plan.
OpenAI Has Realised Its Super App Vision With ChatGPT Work
OpenAI has announced ChatGPT Work, combining the chatbot experience with the capabilities of the software development tool Codex and the browser Atlas on a single platform. This agent can work on a project for hours on the user's behalf and turn a goal into finished work.
DijitalPi's Take
Artificial intelligence is moving from the role of 'assistant answering questions' to 'partner finishing the work'. Being able to hand over tasks that take hours in marketing operations — preparing reports, writing campaign briefs or competitor analysis — to an agent means a serious gain in capacity, for agencies and SMEs working with small teams in particular. Our recommendation: list the repeated tasks that take the most time in your team and test ChatGPT Work as a pilot with one of them.
OpenAI Has Announced a Next-Generation Voice Assistant Experience With GPT-Live
OpenAI has released the GPT-Live-1 and GPT-Live-1 mini models, redesigning ChatGPT's voice interaction features. These models aim to build more natural, more fluent voice communication with the user.
DijitalPi's Take
The voice interface is a candidate to be digital marketing's next big channel. The voice assistant experience becoming more natural opens the way for applications such as voice ordering and voice customer support, in the service and retail sectors in particular; brands' shift from being 'searchable' to being 'speakable' will speed up. Our recommendation: gain an early advantage by testing GPT-Live-based integrations in your customer service and voice search channels, and start a pilot in the health and property sectors in particular, where phone sales are heavy.
Google Is in the News With Migros's Next-Generation Shopping Assistant MAYA AI
Migros has launched MAYA AI, its AI-supported next-generation shopping assistant app. The app aims to offer personalised recommendations by analysing customers' shopping habits.
DijitalPi's Take
Retail giants in Türkiye now position AI not as window dressing but as a core customer experience tool. Migros's step gives a clear signal to its competitors in the e-commerce and retail sector to speed up their personalisation investment, because the customer now expects a recommendation dedicated to them rather than a general campaign. Our recommendation: draw up a roadmap for feeding your own CRM data into AI-supported recommendation engines, and start a personalised recommendation test on a small segment to measure the effect on ROAS (return on ad spend).
The AI Marketing Newsletter is the column where the DijitalPi team comments on the standout developments from the world of artificial intelligence and marketing for brands in Türkiye.
