Welcome to our October Digital Digest!
If any of us were in doubt, the past few months have confirmed it: AI models are stealing your traffic. From AI Overviews to AI Mode, they are fuelling a rise in zero-click searches – but they are also creating more opportunities for brands to build their visibility and authority.
Don’t let headlines like this scare you this October – we’re here to dig deeper and pull out the facts that YOU need to know, all right here in our latest Digital Digest.
The End of The Social Media Age?
The amount of time spent on social media per day peaked in 2022 across most age groups.
It might not feel like it with AI-produced content and convincing deep fakes flooding our feeds, but as a global society we are, on average, spending less time on social media per day compared to 2022. Back then the average number of hours spent on our favourite social apps was 2.5 per day, with the 16 to 24 year old age group spending 3.0 hours every day on social media. However, it was this age group that also began leading the trend of spending less hours on social apps after this 2022 peak.
This age group now spends less time per day on social media compared to 2016-22, with both the 25 to 34 and 35 to 44 age groups showing similar patterns. However, the fact remains that this age group is still spending upwards of 2.5 hours on social apps per day, proving that whilst the amount of time spent on social media may have peaked in 2022, its grip on our attention remains unbreakable.
OpenAI Removes Content Guardrails
ChatGPT to treat ‘adults like adults’, despite internet safeguarding concerns.
We’ve all seen the headlines – ChatGPT to allow mature content! A change in OpenAI policy now means that the AI platform is treating ‘adults like adults’; a proposal that includes reduced restrictions on the models’ ability to generate adult content.
The OpenAI announcement explained that these changes were made due to the existing safeguards being too ‘restrictive’ (an issue that they said resulted in less useful output) but what does this lowering of user guardrails really mean for the platform’s users? The content restrictions were implemented for mental health and user protection purposes, making them an essential safeguarding measure in the mass-production and usage of AI materials. Without them, we take yet another step towards unregulated content production and the dangers this poses, with no clarification yet as to the platforms’ potential for creating harmful or non-consensual images, which remain criminal offences under UK law.
The Fact-Checking Crisis
Where are the verification systems needed to identify errors in AI-generated content?
Calls for AI results to be internally verified, or for there to be a universal fact-checker, have not yet yielded the results needed to combat the ongoing spread of digital misinformation. The recent launch of GPT-5 promised improved accuracy and usefulness in response to the issues of over-compliance, sycophancy and ‘hallucinations’ in its predecessors, but like all AI models, this newest launch ultimately leaves the responsibility of verifying information in the hands of its users.
The closest breakthrough in terms of a fact-checking application came way back in 2024, when researchers at DeepMind developed SAFE – a Search-Augmented Factuality Evaluator – and MIT researchers launched the SymGen tool – an LLM that includes factual citations in all the responses it generates. These AI applications were the first of their kind and despite the increasing reliance on AI across demographics they remain, for now, the only publicly available resources for verifying information delivered by AI platforms.
Introducing Agentic Commerce
OpenAI to offer in-chat shopping for ChatGPT users across the US.
Is there anything that AI models can’t do? Answer queries, summarise reports, produce code, detect cybersecurity threats – and now even help you bag a great deal when shopping online.
OpenAI has recently launched Instant Checkout inside ChatGPT and partnered with Walmart to make in-chat shopping a reality. Users are already able to view product images, details, and links in ChatGPT results, but this new feature doesn’t just help you find products – it helps you buy them too, using its single-item purchasing capabilities.
Interestingly, Instant Checkout’s product results are selected for their relevance and ranked based upon public reviews. This new form of ecommerce is intended to be completely organic and free from any OpenAI partnerships, sponsorship, or paid ads, and represents the next step in agentic commerce – one in which the purchasing decision remains firmly in the users’ hands.
Visibility Vs User Intent
We all know that SEO success is no longer just about ranking – but then what is it about?
In the final quarter of 2025, it’s all about visibility, and the best way to achieve that is by understanding your audience as much as possible. What they’re looking for, what they want, what language feels natural to them; and, most importantly, why they are searching for it.
Once you understand these search behaviours, you can adapt your content in line with them to earn visibility in AI Overviews, structured snippets, and local search results. Content that is structured clearly, answers FAQs, and generates trust will not only achieve this coveted visibility but will also generate user awareness – no matter how well it ranks in organic search.
The Digital Digest
Need help keeping up to date with all the latest developments?
Be the first to know with our monthly Digital Digest, or contact us directly to talk through what these shifts mean for your marketing strategy.
