That is a very broad question, but a good one! I'll answer it in broad terms as I think specific predictions are a dangerous game in this space given the speed of development.
So, some themes for 2026, off the top of my head and in no particular order:
- We'll see less AI seeping into current applications (that's mostly been done now) and more features/capabilities seeping into AI platforms. This builds on the idea that OpenAI are effectively trying to rebuild the internet inside ChatGPT rather than integrate ChatGPT back out into everything. This strategic shift seemed to happen around mid-2025.
- We'll also start to see more AI-native applications arrive in 2026. Most AI development so far has been bolting on AI-features to existing applications and I'll think we'll see this trend replaced by new applications built around AI models from the ground up this year.
- Skills will replace MCP as the predominant way of giving AI agents access to tools. This shifts the idea from different agents that can do different things to one agent that can do lots of things. Skills are also much easier (i.e. less technical) to build, which will lead to faster democratisation.
- OpenAI and maybe Anthropic will IPO (my only specific prediction!).
- We'll learn more about OpenAI's hardware plans, but I don't expect them to launch until. 2027. I also think we'll see something from OpenAI around robotics this year.
- I think we'll see a large brand/celebrity sue an AI company for misrepresentation leading to calls for regulatory intervention and the development of new tools to allow brands/individuals to 'correct' how they're represented by AI models. This might slip to 2027 though.
How do you think AI is going to develop this year?
That is a very broad question, but a good one! I'll answer it in broad terms as I think specific predictions are a dangerous game in this space given the speed of development.
So, some themes for 2026, off the top of my head and in no particular order:
- We'll see less AI seeping into current applications (that's mostly been done now) and more features/capabilities seeping into AI platforms. This builds on the idea that OpenAI are effectively trying to rebuild the internet inside ChatGPT rather than integrate ChatGPT back out into everything. This strategic shift seemed to happen around mid-2025.
- We'll also start to see more AI-native applications arrive in 2026. Most AI development so far has been bolting on AI-features to existing applications and I'll think we'll see this trend replaced by new applications built around AI models from the ground up this year.
- Skills will replace MCP as the predominant way of giving AI agents access to tools. This shifts the idea from different agents that can do different things to one agent that can do lots of things. Skills are also much easier (i.e. less technical) to build, which will lead to faster democratisation.
- OpenAI and maybe Anthropic will IPO (my only specific prediction!).
- We'll learn more about OpenAI's hardware plans, but I don't expect them to launch until. 2027. I also think we'll see something from OpenAI around robotics this year.
- I think we'll see a large brand/celebrity sue an AI company for misrepresentation leading to calls for regulatory intervention and the development of new tools to allow brands/individuals to 'correct' how they're represented by AI models. This might slip to 2027 though.