February

8 events

Editors' summary

Google pushes forward and pulls back at once, OpenAI moves to video

On 8 February, Google renamed Bard to Gemini and launched the paid Gemini Advanced tier. On the 15th it announced that Gemini 1.5 Pro could handle a one-million-token context. On the 21st it released Gemma, with open weights.

The same 15th, OpenAI announced Sora, a video generation model. A coherent one-minute clip drew a strong reaction. It was not released publicly, going only to a limited group.

On the 22nd, Gemini's image generation of people was halted. Prompts asking for historical figures had been returning depictions at odds with the record, and Google said it would turn the feature off until it was fixed.

On the 29th, Elon Musk sued OpenAI and Sam Altman, arguing the company had broken the promises it was founded on.

This block is written by the editors. It is kept separate from the sourced record below.

Record8 events

8
ProductGoogle / Sundar Pichai
Bard becomes Gemini, and Gemini Advanced goes on sale
Google renamed Bard to Gemini and launched Gemini Advanced, a paid tier giving access to the top-end Ultra 1.0 model, included in a $19.99-a-month Google One AI Premium subscription and available in English across 150 countries and territories. Gemini apps for Android and iOS arrived at the same time. Google said that in blind evaluations by third-party raters, Gemini Advanced with Ultra 1.0 was preferred over competing chatbots. The Bard brand, stood up in haste in February 2023, was retired after a year in favor of matching the product name to the model. Priced level with OpenAI's ChatGPT Plus, it marked the point where consumer AI subscriptions became a real contest.
8
PolicyFederal Communications Commission
The FCC confirms that AI-voiced robocalls are illegal
The Federal Communications Commission adopted a declaratory ruling, unanimously, confirming that voices generated by AI count as "artificial" under the Telephone Consumer Protection Act. Under that act, placing automated calls with an artificial voice without prior express consent is illegal. The ruling made no new rule; it stated how far the existing one reached, and it took effect immediately. Chairwoman Jessica Rosenworcel said "bad actors are using AI-generated voices in unsolicited robocalls to extort vulnerable family members, imitate celebrities, and misinform voters," and framed the ruling as giving state attorneys general the tools to pursue them. Three weeks earlier, robocalls using a cloned Biden voice had gone out in New Hampshire.
15
ModelGoogle
Gemini 1.5 Pro reaches a one-million-token context window
Google announced Gemini 1.5 Pro with an experimental one-million-token context window, offered in private preview through Google AI Studio, against a 128,000-token default. The largest window on any publicly available model until then had been 200,000 tokens, in Anthropic's Claude 2.1 — an order-of-magnitude jump. Entire books, whole codebases, or hours of video could be read in a single pass, and long context became a clear point of differentiation for Google. The model used a Mixture-of-Experts (MoE) architecture, running only a subset of its internal "expert" networks for any given input, which Google said made it more efficient to train and serve — the first time Google described a flagship model of its own as MoE. Arriving just two and a half months after Gemini 1.0, it showed how fast Google was moving to close the gap. Context length settled in as a standard axis of competition, with Anthropic reaching a million tokens in 2026.
15
ModelOpenAI / Sam Altman
OpenAI unveils text-to-video model Sora
OpenAI unveiled Sora, a model that generates videos of up to one minute from text prompts. Sora produces high-fidelity footage that closely follows instructions, and can also animate still images and extend existing videos. At announcement it was not generally available, limited to red-team safety evaluation and a small group of creators, but the quality of the demo videos drew widespread attention and accelerated the race in AI video generation.
21
ModelGoogle
Google releases Gemma, its open-weight models
Google released Gemma, a family of lightweight open-weight models built, it said, from the same research and technology as Gemini. It started with two sizes, 2B and 7B parameters — the 7B aimed at consumer-grade GPUs and TPUs, the 2B at CPUs and on-device use — both licensed for commercial as well as research use. Google was careful to call them "open models," open-weight with terms of use attached rather than open source. It marked Google's entry into the open-weight territory Meta's Llama had held, and settled the pattern common across the industry: keep the frontier model closed, open up the small ones.
22
ProductGoogle / Sundar Pichai / Prabhakar Raghavan
Google halts Gemini's image generation of people
Google paused Gemini's generation of images of people after a run of historically inaccurate output: racially diverse German soldiers of the Second World War, a Black female pope, multiethnic American Founding Fathers. Diversity tuning intended to counteract bias in the training data had been applied without distinguishing prompts where it was inappropriate, and it overcorrected. SVP Prabhakar Raghavan wrote the next day that the model had "overcompensated" and been "over-conservative"; on 27 February Sundar Pichai told employees in a memo that the images were "completely unacceptable" and said the company would run a structural review of what went wrong. The episode exposed how an intervention meant to correct bias can introduce a different inaccuracy, and it became politically charged. Generation of people resumed in August that year.
27
ProductMicrosoft
GitHub Copilot Enterprise becomes generally available
GitHub made Copilot Enterprise generally available at $39 per user per month. On top of the Business tier, its central feature is the ability to draw on an organisation's own private code and documentation, so answers follow the company's own practices, with tuning against an existing codebase. GitHub pitched it as a way for developers new to a team to ask how things are done there.
29
PolicyOpenAI / Microsoft / Elon Musk / Sam Altman
Elon Musk sues OpenAI and Sam Altman
Elon Musk, a co-founder who left OpenAI in 2018, sued OpenAI, Sam Altman, and Greg Brockman in San Francisco Superior Court. He alleged that the company had abandoned its founding agreement to develop AI for the benefit of humanity, prioritizing profit through its relationship with Microsoft. On March 5, OpenAI pushed back by publishing early emails suggesting Musk himself had acknowledged the need for a for-profit structure and large-scale funding.

↑ 2024