A week in Generative AI: Exodus, Burnout & Deep Research
News for the week ending 15th February 2026
This week’s newsletter is a bit more personal. Two AI-related stories caught my attention this week that directly challenge the productivity narrative we keep hearing. One shows people losing jobs to AI efficiencies. The other reveals what’s happening to those who keep their jobs: they’re not working less, they’re working more. As someone who uses AI heavily and has dealt with burnout, both hit uncomfortably close to home.
In more typical AI-related news though, there was just one thing that stood out, which was an update from OpenAI to Deep Research, a big feature I think is mostly slept on.
In Web 4.0 news, OpenAI started testing ads in ChatGPT, and Google previewed WebMCP, a new protocol for AI agent interactions. In Ethics news OpenAI disbanded its mission alignment team and Anthropic said it’ll try to keep its data centers from raising electricity costs.
I also highly recommend the Long Read from the NY Times - I Left My Job at OpenAI. Putting Ads on ChatGPT Was the Last Straw.
UK ad agencies undergo their biggest exodus of staff as AI threatens industry
It’s fair to say that the ad industry has had a tough few weeks. There have been big losses reported, as well as a sell-off fuelled by a wider sell-off in tech stocks following the announcement of Claude Cowork. There’s also been lots of agency consolidation and marketing budgets have flatlined.
All of this has been building for a while, and has resulted in advertising agency staff in the UK shrinking by 7% YoY with creative agencies taking a bigger hit with a 14% drop.
It’s very easy to blame this on AI, but the reality is that the advertising industry is going through huge structural changes at the moment. This is mostly driven by how the advertising ecosystem has been changing over the last 5-10 years with shifts towards performance marketing, the rise of influencers, and the dominance of the large digital platforms.
AI will absolutely drive a lot of change throughout the advertising industry, and sadly this is being looked at very one dimensionally at the moment. This is being driven by an incredibly competitive agency market that is driving down prices which agencies are hoping to recoup in the medium term by finding AI-driven efficiencies, even though these efficiencies are yet to be realised. AI technology is still maturing, the infrastructure needed to fully leverage it needs building out, and new skills and ways of working need to be established. The human cost of this transition is already being felt, particularly among younger workers and recent graduates who are finding it increasingly difficult to break into the industry.
My hope is that AI will open up a huge amount of opportunity alongside the change that it will drive. AI platforms are now used by c.1.5bn people every week worldwide and have matured into a viable marketing channel in their own right. Helping brands navigate this exciting and emerging space will require a lot of knowledge and expertise and advertising agencies should be well placed to provide this if they’re able to make the right investments and not just focus on cost cutting and efficiencies.
The first signs of burnout are coming from the people who embrace AI the most
While the previous article looked at job losses in the advertising industry, recent research published in Harvard Business Review reveals what’s already happening to those who retain their jobs. As someone who has suffered from burnout and is a big user of AI, this particularly resonated with me.
What this study found was that workers are adopting AI because it feels empowering, but this leads to taking on more tasks and doing more multitasking. The researchers describe this as a self-reinforcing cycle: AI accelerates tasks, which raises expectations for speed, which makes workers more reliant on AI, which expands the scope of what they attempt, which further increases workload. Workers found themselves taking on tasks that used to require specialists, prompting AI during lunch breaks, and juggling multiple AI-assisted workflows simultaneously, all of which felt productive in the moment but accumulated unsustainably.
This directly undermines the efficiency logic driving job cuts in the ad industry. When agencies reduce headcount expecting AI to fill the gap, they’re not creating sustainable efficiency, they’re redistributing work to fewer people who will work at higher intensity. The productivity gains are real in spreadsheet terms, but they’re achieved by extracting more from those who remain.
The danger is that the cognitive fatigue, burnout, and weakened decision-making that follows are lagging indicators - it will take time for the signs to appear. The question is whether organisations will recognise this dynamic before the initial productivity surge gives way to the inevitable crash. Because right now, most are doubling down on the efficiency narrative while the teams using AI every day are quietly drowning under an ever-expanding workload.
OpenAI Updates Deep Research
OpenAI released some nice updates to Depp Research in ChatGPT this week allowing users to specify sources to search and to be able to update research with new sections and analysis after the initial run.
I think Deep Research is one of those ‘slept on’ capabilities in the AI world - hugely useful and powerful when you use it, but I don’t think many people do.
Web 4.0
Google previews WebMCP, a new protocol for AI agent interactions
Anthropic’s Super Bowl ads mocking AI with ads helped push Claude’s app into the top 10
AI Ethics News
EU Issues Warning to Meta Over Closing Off WhatsApp to AI Rivals
OpenAI disbands mission alignment team, which focused on ‘safe’ and ‘trustworthy’ AI development
Anthropic says it’ll try to keep its data centers from raising electricity costs
Researchers Jailbreak ChatGPT to Find Out Which State Has the Laziest People
Long Reads
NY Times - I Left My Job at OpenAI. Putting Ads on ChatGPT Was the Last Straw
Harvard Business Review - AI Doesn’t Reduce Work—It Intensifies It
“The future is already here, it’s just not evenly distributed.“
William Gibson