OpenAI, the company behind ChatGPT, has quietly shifted its millions of free users to a cheaper, less powerful version of its artificial intelligence, a move framed as user choice but widely seen as a significant cost-saving measure. For Kenyans, who rank as the world's top users of ChatGPT, this change could impact the quality of a tool that has become vital for work, education, and small businesses.

The change means that non-paying users will now default to GPT-5.2 Instant, a model designed for speed and everyday tasks, instead of having some complex queries automatically routed to the more powerful 'Thinking' model. While users can still manually switch to the 'Thinking' model for more demanding tasks, the company is banking on the likelihood that most will stick with the default.

Why This Matters in Kenya

This is not just a technical switch; it has practical consequences for the Kenyan dinner table. Recent data shows that a staggering 42.1% of Kenyan internet users over 16 use ChatGPT, the highest rate in the world. Students use it for research, entrepreneurs for drafting business plans, and developers for writing code. Access to a high-quality AI for free has been a significant leveller.

The core of the change lies in the difference between the two models:

• GPT-5.2 Instant: Described by OpenAI as a "powerful workhorse for everyday work and learning," it prioritizes speed.
• GPT-5.2 Thinking: This model is built for deeper, more complex reasoning, such as solving difficult work tasks, coding, and scientific research.

By making the less capable model the default, OpenAI reduces its operational costs, as processing complex queries on advanced models is resource-intensive. The move affects not only free users but also subscribers to the lower-cost 'Go' plan, which costs $5 (approx. KES 645) per month.

A Question of Transparency and Trust

OpenAI has presented the update as a way to "maximize choice." This follows previous user frustration over the platform automatically switching models without clear consent. However, the financial incentive is hard to ignore. The company has been actively exploring strategies to manage and reduce the immense costs associated with running its powerful AI systems.

A critical concern is how the less powerful default model will handle sensitive queries. Previously, OpenAI noted it would route queries showing signs of mental distress to its more advanced models for better responses. While the company claims GPT-5.2 Instant is now better equipped for these situations, the change removes an automatic safety net.

As Kenya continues its rapid adoption of AI, with the government launching a National AI Strategy in March 2025, the reliability of freely accessible tools like ChatGPT remains crucial. This shift by OpenAI serves as a stark reminder that the future of 'free' technology depends on a delicate balance between corporate costs and user value.