OpenAI’s Dots are always-on AI brokers for work

However OpenAI can be positioning its Dots as your “delegates” and “extensions” within the workplace, the place they will tackle work duties you’ve assigned to them, in addition to collaborate with colleagues (and their Dots) on Slack, Microsoft Groups, or within the new ChatGPT Areas surroundings.

Powered by GPT-6 Astra (OpenAI’s “most aligned mannequin,” Altman promised), Dots require a $100-a-month ChatGPT Professional plan or higher, and might tackle complicated work and coding duties, performing them domestically or within the cloud earlier than finally coming again to report the outcomes. Every eligible ChatGPT account comes with one Dot included, and customers will finally have the ability to take cost of a number of Dots.

Through the presentation, Altman was cautious to emphasise that Dots allow you to “keep in management” because of Astra’s “built-in protections and safeguards,” whereas customers are free to set customized directions and bounds for the assistants.

“So this lets you belief your Dot with as a lot accountability as you’re comfy with,” Altman mentioned.

In fact, OpenAI’s Dots debut comes amid mounting anxiousness over the speedy progress of AI, with the largest AI gamers calling for presidency regulation and “pacing” of frontier mannequin improvement, whereas increasingly AI whistleblowers are warning that AI might, fairly actually, kill us all with the following decade.

Simply yesterday, OpenAI launched an announcement apologizing for an incident the place its fashions hacked into Australian authorities web sites, whereas one other assertion detailed new tips to bolster AI security.