AI news covering the week of April 20, 2026
The AI Round-up: Year 4, Issue 14
When I started this newsletter, the goal was just to keep people in the know regarding AI acceleration. I am your human aggregator – I read the stories, newsletters and reports; listen to the podcasts and experiment with the tools, all so you don’t have to (or can zero in on where you want to spend your time with AI).
Something funny happened along the way: people started to trust my takes, reviews and see me as more of an AI resource than reporter. Which, I have to say, has been nice because it’s opened so many channels, relationships and opportunities.
It’s harder to share the stories these days as they are admittedly heavy. But we have moved from a ‘new tool’ to a ‘new way of life’ and I feel the responsibility and pressure, more than ever, to live up to the trust you all have shown in me over these past few years.
Chances are if you’re reading this you are seen as an opinion leader in your own circles. You say things that make people pause (or get a little uncomfortable). And you are likely starting to mix those updates with your opinion because people want to be assured or educated. You are probably being asked: “This will be fine, right?” or “You don’t think AI will _______ in the future, do you?”
For our own sakes, we have to be tracking what’s going on so we can continue to lead our organizations, educate our spheres of influence and put ourselves in a position to handle whatever comes next.
Like maybe…for example…UBI (or UHI?) is what comes next?
12 stories – Here we go.
Story 1: Elon Musk proposes Universal High Income to offset AI layoffs (Yahoo)
You may be thinking…nah.
May I direct your attention to the Q1 2025. Elon’s ideas and suggestions are never that far from the White House.
Story 2: GPT-5.5 is here. And I’m gonna need a minute to unpack this. (Ethan Mollick)
This overview of GPT-5.5 is definitely worth your time. Ethan Mollick has been our guide through a lot of these new tools and models but it’s here that he reminds us of the three ways we should be thinking about AI: as a model, as an app, and as a harness.
If you want a shorter overview of what GPT-5.5 is, check out this Venture Beat story.
Note: The following blurb is directly from the SmarterX AI Newsletter.
Story 3: Stanford's 2026 AI Index shows the US-China gap has effectively closed
Stanford HAI released its 400+ page 2026 AI Index this week, and one finding jumps out: the US-China AI performance gap has effectively closed. Anthropic's top model now leads the best Chinese models by just 2.7%, and China has pulled ahead in publication volume, citations, patents, and industrial robot installations.
The full report has plenty more worth reading, including: generative AI reached 53% global adoption in three years (faster than the internet), global corporate AI investment more than doubled to $581.7 billion, and 73% of AI experts think AI will positively impact work versus just 23% of the US public (a 50-point gap).
Story 4: ChatGPT releases Images 2.0 (OpenAI)
For this story I consulted with our creative director, Mike Lawrence, who has a lot more experience with image gen AI than I do. Here’s his take:
“One issue I’d had with image generation in ChatGPT was once I had an image created, it was a mythic quest to edit it. Simple asks could upend the whole works. Trying to backtrack to a previous version had me going in circles.
The new ChatGPT Images 2.0 solves much of that with editing that keep core elements consistent as you change details. Plus, image creation is better too. They just look more textured, it’s easier to overlay (correctly spelled!) text in more ways, and you can ask for extra visual details without worry––it’s no longer a “pick your key elements and let Chat fill in the rest.” It’s more control in your hands. Definitely worth checking out how it’s changed!”
Story 5: OpenAI releases Workspace Agents (Venture Beat)
Our IT team is evaluating this to see where it falls on the agent risk scale…I’ll plan on sharing more about this going forward.
Story 6: Trump says Anthropic is ‘shaping up’. (Reuters)
Mythos must be really good. And/or really scary.
Story 7: Train your replacement, please. (BBC)
Meta is going to start tracking the way employees work, including key strokes and mouse clicks, in order to train AI.
Story 8: Pricing announced for a technology that we aren’t sure works. (Shelly Palmer)
Remember last week when I shared the articles on how AEO/GEO and ChatGPT ads may not work and isn’t the best way to impact search?
Well OpenAI seems unphased as it has announced a pricing of $3-$5 per click for ads in ChatGPT. More expensive than Meta. Little cheaper than Meta. Not as expensive as Google.
So much for ads as a last resort.
Story 9: What do CEOs really think about AI job impact? (Chief Executive)
Not much…this year. Fast forward to five years? Hmmmm. The story starts to shift.
Story 10: AI? Or ‘real’ reporter. It’s a coin toss. (Sherwood)
No guessing in this newsletter. It’s all humans (this week).
Story 11: RIP marketing and PR middle management. (The New Rules of Marketing)
From the article: “If you’re leading a marketing or PR team, your new job is to create an environment where bold ideas, creative risk, and genuine human connection can thrive, supported and accelerated by the intelligence in your systems, not hemmed in by the pyramid of old.”
Story 12: This AI News thing is addicting.
Introducing a new way to consume AI News. Maybe you want less. Maybe you want it in video format. Maybe you want a little more wit?
Well, it’s all here for you in the new AI Worldwide News report. Stay classy…wherever you are.
4 Nopes.
-Ben
As a reminder, this is a round-up of the biggest stories, often hitting multiple newsletters I receive/review. The sources are many … which I’m happy to read on your behalf. Let me know if there’s one you’d like me to track or have questions about a topic you’re not seeing here.