Sam Altman wants to hear from you on what OpenAI should build next year. Here are 3 ideas that have caught his eye so far.

• Sam Altman asked his followers on X for ideas to improve OpenAI in 2025.
• Altman responded to suggestions about creating family accounts and improving video features.
• OpenAI had a rocky year following new litigation from competitor Elon Musk.

Sam Altman wants your suggestions for improving OpenAI in the new year.

On Christmas Eve, OpenAI’s CEO opened the floor to his followers on X with a simple question: “what would you like openai to build/fix in 2025?”

Altman’s post comes toward the end of a tumultuous year for OpenAI. Elon Musk, who cofounded the company but left the board in 2018, filed a lawsuit against Altman and OpenAI in February. Musk withdrew the lawsuit in June but filed a new lawsuit in August in which he argued that OpenAI executives “deceived” him into cofounding the company.

Amid its legal challenges, OpenAI raised $6.6 billion in funding in October — a record-high funding round for tech in Silicon Valley — pushing the company’s valuation to $157 billion.

Now, Altman is looking at what’s in store for the new year, and he’s crowdsourcing ideas on how to continue advancing and improving not just ChatGPT but all of the company’s existing and emerging products.

Here are three ideas Altman appears to be eyeing so far.

Family accounts

One X follower responded to Altman’s post with a suggestion to allow OpenAI users to create accounts for families that include guardrails, allowing kids to use the platform safely.

“Let their curiosity take off, but within reasonable limits, as determined by the parent,” the user wrote, to which Altman responded: “this is a good idea!”

OpenAI’s website states that ChatGPT isn’t meant for children under 13 and requires parental consent for children ages 13 to 18 to use the platform, but does not actively verify the consent.

Improvements to voice chats

ChatGPT developed a voice feature that allows users to speak directly to the chatbot. One of Altman’s followers suggested that the feature be improved because, as of now, per the user, “you can’t just silently think for 5 seconds without being interrupted.”

Altman wrote that that’s a “good point.” Users have struggled with the feature since it was rolled out, with some having reported bugs and misinterpretations from the chatbot.

Better video generation

Multiple users urged Altman to improve Sora, OpenAI’s text-to-video model. Sora allows users to describe a video they want the platform to create with a written prompt, which Sora will generate based on the text.

Sora officially launched on December 9 after being piloted to a limited group of creators in February. Rohan Sahai, Sora’s product lead, said at the launch that the product would begin more conservatively because its team wants to prevent illegal activity like copyright violations while allowing for creative expression.

OpenAI did not immediately respond to a request for comment from Business Insider.