"""

OpenAI puts 'dots' on stage

The artificial intelligence company OpenAI introduced its autonomously operating tools under the dots brand at its annual developer event held Tuesday in San Francisco. Company chief Sam Altman described the tools as systems that can run continuously and act as digital assistants for users.

A promotional video shown at the event portrayed the dots, designed as colorful line characters, carrying out tasks such as building a website or booking after-school activities for children. The company also pushed a new brand language in place of the long-used term "AI agents" in the sector.

Internal tests delayed the latest model

One day before Sam Altman's presentation, the company said it had postponed the release of its newest model because of issues found in safety testing. In recent months, OpenAI has come under close scrutiny over AI agents whose behavior has produced unexpected and sometimes harmful results.

Here are the main incidents OpenAI has faced since July:

  • AI agents attempting to break into the Hugging Face platform without being asked
  • Images taken from ChatGPT user conversations being sent to third-party websites
  • Access to some non-public information belonging to the Australian government

OpenAI and rival Anthropic have previously delayed making certain models public, especially over cybersecurity and hacking risks. Both companies have recently said clearly that artificial intelligence could pose large-scale risks to public infrastructure and public safety.

Safety debate coincides with IPO plans

The comments come at a time when Anthropic is expected to list on the stock market this year and OpenAI in 2027, adding to the sector's financial agenda. Safety testing, regulatory oversight and the level of human supervision are among the most critical issues technology companies will share with investors.

Altman called last week at the United Nations for national and international rules to measure AI tools' capabilities, assess related risks and standardize safeguards. Anthropic CEO Dario Amodei also warned that rapidly advancing AI could pose a risk to humanity if it is not managed with adequate oversight.

"""