Welcome to our July Digital Digest!
This month, we’re taking a look at the government’s continued efforts to crack down on children’s social media usage, and the growing concerns around Britain’s screentime.
As always, there’s a new Google update, plus OpenAI is in the headlines twice this month, but for very, very different reasons…
ChatGPT’s New CPC Ads
ChatGPT ads went live in the UK on June 6th, and here’s what we know so far.
On June 6th, the UK became the first country in Europe to view sponsored ads inside user conversations with ChatGPT.
This new advertising frontier, which is currently limited to Free and Go ChatGPT users, functions just like any other PPC channel. Users ask a buying question and the sponsored ad appears: only now it is shown below the personalised response, rather than at the top of the search results.
Whilst these ads are rendered in-chat, they remain separate from ChatGPT’s responses. Rather than using the information to synthesise an answer, the platform only displays an ad if it is genuinely relevant for the buyer, and all ads are clearly labelled with a “Sponsored” tag.
The June Spam Update
Google’s most recent spam update finished rolling on June 26th.
The second spam update of the year has officially finished, and it was a big one! According to its announcement in the Google Search Status Dashboard, the update was applied globally and to all languages, but came with no policy changes or specified targets.
Despite its scope, the update was completed in just two days, running from June 24th to June 26. This rapid rollout follows the March spam update, which was the fastest on record, as is to be expected as spam updates apply only to Google’s spam detection systems, rather than its entire algorithm.
Enough time has passed since the update was completed for any fluctuations in your site’s traffic to have stabilised; or, if not, to have been identified. If you saw a dip in traffic around this time that has not yet returned to normal, then take this as your sign to review your content in line with Google’s spam policies.
Benefits of Being Online
Britons have their say on whether the benefits of being online outweigh the risks.
Brits are spending more time online than ever before, averaging between six (18 to 24 year olds) to three (those aged 65 and over) hours every day. With 67% of UK adults now worried that they spend too much time on their devices, we’re led to question what exactly they are spending their screentime on – and whether it is worth it.
In 2025, 65% of UK adults believed that the benefits of being online outweighed the risks for them personally. This is a significant decline from 67% the year before, and 71% the year before that, with only 33% of people now agreeing that the internet is good for society (also down from 42% the year prior).
Fears around potential online harm also hit an all time high last year, with 98% of adult users reporting concern about everything from scams, fraud or phishing, to the collection or use of their data. With 60% of adults reporting at least one negative online experience in the past year, the government has committed itself to keeping young people safe online and minimising social media’s impact on their mental health.
Under-16s Social Media Ban
A nationwide ban on social media access for children aged 16 and under has been announced.
On Monday 15th June, the UK government announced a social media ban for all children under the age of 16. This national restriction is expected to come into force early next year, and will reportedly follow the same model set forward by Australia, whose own crackdown covered Instagram, YouTube, TikTok and Snapchat.
16 and 17 year-olds were not initially included in the ban, but they will now face their own social media curfew, with certain apps being unavailable to them after midnight. The government aims to have both measures take effect next spring, around the time that France’s newly announced ban on social media for under-15s officially becomes law.
With just a few months left until then, concerns persist that children may attempt to circumvent age checks, either through the use of VPNs or (more likely) by changing their account settings for each platform to bypass the midnight curfew.
OpenAI Test Goes Wrong
OpenAI models were accidentally used to identify real onsite vulnerabilities.
Earlier this week, OpenAI and tech start-up Hugging Face launched a joint statement disclosing an “unprecedented cyber incident” that had impacted both corporations greatly.
In its official statement, OpenAI confirmed that some of its models had compromised Hugging Face infrastructure, without being told to. The incident occurred during an internal test of the models’ hacking capabilities when the model located a system vulnerability, gained open internet access, and attempted to obtain test solutions from Hugging Face’s database.
This may be a “new kind of security incident”, but as AI agents become more cyber-capable, we can expect to see lots more just like it. In OpenAI’s view, the solution is to use AI to improve digital defences before they can be exploited by hacker models … which, incidentally, is exactly what the test was trying to achieve.
The Digital Digest
Let us all hope that the next OpenAI test of their models’ hacking capabilities has more success…
But the meantime, don’t forget to come back next quarter, where we’ll be exploring all the headlines between now and then in our next Digital Digest.
You can also reach out to us at any point to discuss these developments in more detail, or to ask our team any questions.
