The journey of Artificial Intelligence (AI) spans several decades, marked by the dedication of tech leaders and research institutions. Yet, it was only when OpenAI rolled out ChatGPT in November 2022 that AI caught the public’s eye, rapidly accumulating over 1 million users within just five days.
Today, that user base has surged tremendously, positioning ChatGPT as the go-to reference for many when discussing AI. OpenAI is in a constant cycle of advancements, launching new models and features at a brisk pace, with 2026 poised to be a landmark year.
Several noteworthy changes are on the horizon for ChatGPT, following the recent introductions of personality presets and ChatGPT Health in late 2025 and early 2026, respectively. It’s essential to note that the upcoming details stem from OpenAI’s public disclosures, and further innovations likely remain under wraps. For those closely monitoring developments, it’s hoped that some of the much-needed features for ChatGPT are also in the works for 2026.
GPT-5.2 to Replace Previous Models
For those keeping abreast of AI advancements, it’s common knowledge that companies like OpenAI regularly introduce new iterations. Each updated model is designed to enhance intelligence, speed, and capability. Typically, older models get phased out shortly after their successors are unveiled.
On February 13, 2026, OpenAI plans to retire six existing models at once, including GPT-5 (both Instant and Thinking), GPT-4o, GPT-4.1, GPT-4.1 mini, and OpenAI o4-mini. While these models will no longer feature in ChatGPT, they will still be accessible via OpenAI’s API. Following this date, GPT-5.2 will serve as the default model for all ChatGPT users.
Notably, this will mark the second attempt at deprecating GPT-4o. The initial effort coincided with the rollout of GPT-5 but had to be reversed as users, particularly Pro and Plus subscribers, voiced their preference for GPT-4o’s conversational warmth. OpenAI has taken this feedback into account, asserting that they have improved its shortcomings, leading to a significant portion of users now favoring GPT-5.2, leaving only 0.1% still opting for GPT-4o.
Introduction of Advertisements
Advertisements are generally unpopular, and OpenAI’s CEO, Sam Altman, expressed his disdain for mixing AI with ads back in October 2024, stating that it would be a “last resort.” However, approximately 20 months later, that last resort has become a reality. In mid-January 2026, OpenAI announced its plans to initiate ad testing within the United States via social media and press releases.
This announcement coincided with the launch of the new “ChatGPT Go” tier in the U.S., a subscription service priced at $8/month, positioned between the free service and the $20/month Plus subscription. Both free users and Go subscribers will encounter ads, while Plus, Pro, Business, and Enterprise accounts will maintain an ad-free experience. OpenAI assures that advertising will not manipulate ChatGPT’s responses, and any ads displayed will be clearly marked and kept distinct from the content.
Moreover, user interactions and data are safeguarded and not for sale to advertisers. OpenAI reiterated its commitment to user trust and experience over profit, although the move towards advertisements seems to contradict that assertion.
To mitigate potential backlash, OpenAI stated that ads will not be displayed alongside sensitive topics, such as mental health and politics. Currently, testing is limited to logged-in adult users in the U.S. If you have yet to see ads in ChatGPT while using the free or Go subscription, you might be fortunate… for now.
Implementation of Age-Aware Safeguards
With the increasing accessibility of the internet, the presence of AI chatbots that can answer a myriad of questions has raised significant concerns regarding child safety online. OpenAI is aware of these challenges and is beginning to implement an “age prediction” feature worldwide.
This feature is designed to estimate a user’s age to ascertain if they are above or below 18 years. Per OpenAI’s support documentation, this system uses various signals associated with a user’s account, such as common discussion topics and typical usage times.
For instance, if a user typically engages with ChatGPT after school hours or late in the evenings—times often linked to younger users—the system may classify that account as potentially belonging to a minor. Similarly, frequent discourse on topics appealing to younger audiences could trigger a similar assessment.
When it’s established that a user is under 18, additional safety settings will be activated. This will ensure that sensitive subjects, including graphic violence, sexual content, self-harm, and negative body image portrayals, are handled delicately. However, there is also the possibility that users may be misidentified as minors, in which case they can manually verify their age through the ChatGPT website.
Responses from ChatGPT to Include Visual Elements
Traditionally, AI-generated responses, including those from ChatGPT, can often veer toward being monotonous. Many users desire succinct answers rather than lengthy explanations. The developers at OpenAI seem to have recognized this demand and are already making strides toward improvement, although specifics remain somewhat unclear.
On January 30, 2026, OpenAI’s updates indicated that ChatGPT’s responses are now “more visual and easy to scan.” The enhanced responses for common inquiries may feature “at-a-glance visuals,” while key figures, locations, products, and concepts will be highlighted within the text. Users can then click these highlights to access a side panel with essential details and reliable sources.
Details remain sparse about how this will materialize in practice, but early speculation suggests it could resemble Google’s knowledge panels. The feature has started to roll out globally across web ChatGPT as well as its mobile applications on Android and iOS.
