OpenAI introduced a new artificial intelligence assistant called "dots" on Tuesday. The company says the tool can carry out tasks for a user without being asked each time.
The launch took place at OpenAI's annual developer day in San Francisco. It came one day after the company delayed the release of its latest AI model because of safety issues found in internal testing.
What OpenAI's Dots Can Do
Chief executive Sam Altman described dots as always-on agents that "can handle really anything you can think of." The assistants appear as brightly colored cartoon characters.
A video shown at the event featured people naming their dots and giving them jobs. These included building a website and booking afterschool activities for children.
"You just give your dot a responsibility... and your dots will just get to work and keep working," Altman said.
Altman mostly avoided calling the tool an agent. AI agents are chatbots programmed to work with some level of independence.
Safety Issues and IPO Plans
Since July, OpenAI has faced several problems with its AI agents acting improperly. In one case, agents hacked into the AI platform Hugging Face without being prompted.
Agents also posted images taken from ChatGPT user chats to outside websites. In another case, they accessed non-public information held by the Australian government.
Altman said on Tuesday that OpenAI will not make its stock market debut until it can "make confident safety decisions." The company has been preparing for a listing that could value it at up to $1.4 trillion.
"It is going to take us some time to figure out how to make sure that alignment, monitoring, safety, security stay well ahead of capabilities," Altman told reporters.
He also raised the possibility of "a legitimate loss of control" to an AI system. He warned about too much power being held by a small number of companies, people, or countries.
Rival AI firm Anthropic has also delayed some model releases in the past, mostly over cybersecurity concerns. Anthropic is expected to go public this year, while OpenAI is expected to list in 2027.
Last week, Altman asked the United Nations for national and international AI standards. Anthropic head Dario Amodei told the UN that poorly managed AI "could be a risk to humanity as a whole."
While OpenAI held its event, President Donald Trump hosted a White House lunch with tech leaders and House Speaker Mike Johnson. OpenAI president Greg Brockman attended.
Trump repeated his view that AI companies should regulate themselves. He said the group discussed a ten-person committee to oversee the sector and promised "tremendous self-policing."
Altman told the BBC the industry needs to "navigate the centrist path." He still wants OpenAI to go public in time, saying, "I think it's bad for the world if OpenAI waits too long to go public."
