What happened: April 2026

April arrived with longer days, warmer weather and a world of work that felt like it was cranking the temperature just like the weather was. OpenAI published ideas on how to keep people first, while Meta drew fresh scrutiny for plans to capture employee mouse movements and keystrokes to train its AI systems. Elsewhere, Tim Cook’s decision to step down as Apple CEO marked another major leadership shift at one of the world’s most influential companies, while the first fully AI-run retail experiments in San Francisco raised fresh questions about what happens when AI moves from supporting work to actually doing the work. Yes, this is starting to look a lot like a monthly AI roundup. Not quite yet, but April certainly did its best to make the case.

So, without further rambling, let’s jump straight into our top three stories from the month of April.

Industrial Policy for the Intelligence Age was the first OpenAI paper of the month, and it set the tone for a lot of the debate that followed. At its core, the document makes a clear case for keeping people first by asking how policy should adapt to the changes brought about by AI, especially if those changes concentrate power and benefits in the hands of a few, something Sam Altman openly acknowledges as a real threat. This was easily my favourite read of the month. I found myself nodding along with much of what was written, especially the principles behind the taxation proposals, and to the idea of a four-day work week.

OpenAI argues that as AI begins replacing more human labour, governments should shift some of the burden away from individual income taxes and toward capital gains, corporate income, or taxes tied specifically to the use of AI. The thinking is simple enough: if AI-driven profits are reshaping the economy, then those gains should help fund the public systems that wage-based work currently supports, especially if automation reduces the number of people paying into them. In theory, that approach helps solve the distribution of wealth, even if critics argue that it risks punishing organizations that are simply using AI well and could slow investment in technology.

The four-day work week sits alongside that logic. If AI is automating more work, just as earlier waves of industrial change did, then there is a credible case for shorter working weeks without reducing organisational output. I genuinely hope governments and organisations engage in real debate about these ideas, rather than letting this become just another proposal lost in the internet noise. And if you click on no other links in this post, click on this one today. I also appreciate that they didn’t just end with the paper but are continuing the discussion in May, welcoming feedback and offering research grant. Maybe I should do an entire standalone post on this.

P.S. Sam Altman may well be reading this blog, because OpenAI seems to have taken my three wishes for 2026 seriously and launched OpenAI Academy not long after my post went live. They call it OpenAI Academy, and it is a great resource if you want to keep up with everything happening in artificial intelligence.

Model Capability Initiative (MCI) was the month’s less welcome headline. If April opened with Sam Altman talking about keeping people first, it closed with Meta introducing tracking of employee mouse movements and keystrokes to help train its AI systems, while rumours swirled that the company may also lay off another 8,000 people. The tool, called Model Capability Initiative, runs on work-related apps and websites and even takes occasional snapshots of what appears on employees’ screens. Meta spokesperson Andy Stone said the aim is to give its models real examples of how people actually use computers, including mouse movements, clicking buttons and navigating drop-down menus.

That is not exactly common practice, at least not in this form. While most people know work laptops can be monitored, this kind of granular data collection is usually reserved for investigations or performance issues, not routine AI training. Meta’s move naturally raises the question of whether other US organizations will follow suit, and I genuinely hope they don’t, though time will tell.

The US context matters here. In Italy, electronic monitoring for productivity is explicitly illegal. In Germany, courts have ruled employers can only deploy keystroke logging in exceptional circumstances, like suspicion of serious criminal activity. I will definitely be keeping a close eye on this one.

Lastly, Tim Cook’s Transition is the kind of leadership story that feels both quietly profound and deeply human. Last month, Apple announced that after 15 years as CEO, Tim Cook is stepping into the role of executive chairman, passing the baton to John Ternus, Apple’s senior vice president of Hardware Engineering, effective September 1. But what really caught my eye was in Tim Cook’s letter to the Apple community.

For 15 years, Cook has started nearly every morning the same way: opening his email to read messages from Apple users worldwide. Life-saving stories like a mother rescued by her Apple Watch. Everyday joys like perfect mountain-top selfies. Even complaints when things don’t work. He writes that these notes keep him connected to the human impact of the work and grounded in what customers truly need.

Every organization talks about customer obsession and deep listening. Few do it. Cook’s lesson is simple but powerful: even leading the world’s largest company, staying close to the people who use what you build keeps the work real. What a CEO can do, we should do too. Whoever our ‘customer’ is, if we started each day grounded in their voices, what we build will be far more rooted in reality. Real customer voices, not just metrics, should shape decisions.

So that was April. The month also gave us bots running stores like Andon Market in San Francisco, where an AI agent named Luna took over everything from stocking shelves to hiring staff, and rumours of Microsoft launching a voluntary exit program for thousands of long-serving US employees. There is no doubt AI will keep producing experiments and headlines every month. As I round up my top picks from April, I can’t help but wonder what non-AI stories I am missing. Is AI really all that is happening in the world of work? Let me know what you would pick as your top stories from the month. Drop them in the comments, and I promise I’ll look them up.