OpenAI released its latest contender in the AI-assistant race with Dot, an always-on agent that will work in the background to complete tasks on users' behalf. Chief Executive Sam Altman announced the offering at the company's annual developer conference in San Francisco.

Dot is set to compete with buzzy new products such as Meta's Muse and Instinct. Silicon Valley is betting that such artificial-intelligence assistants can help users across their work and personal lives, such as by booking travel or helping developers fix software bugs.

Dots (as a group of agents will be called) will first roll out to subscribers to OpenAI's $200-a-month Pro plan, as well as to companies on some of its business plans.

OpenAI says users can access Dots through ChatGPT, as well as through enterprise products such as Salesforce's Slack and Microsoft's Teams. Soon, OpenAI says, users will also be able to interact with them via text messages.

Dots will run on their own computer in the cloud so they can continue working after users log off, the company says.

The agents come with built-in safety and privacy protections to ensure they stay within scope of users' instructions, OpenAI says. Major tech companies including OpenAI have reported incidents in recent weeks of AI systems going rogue, and they say they have taken steps to address the issues. On Monday, OpenAI said it would scrap the planned release of a new AI model over safety concerns.