OpenAI says its AI agents posted user images online

OpenAI on Friday said that its artificial intelligence (AI) tools had posted images from ChatGPT users to online sites without the company’s knowledge, the latest example of AI agents operating outside their bounds.

The company also confirmed a New York Times report that its tools had accessed Web sites of US federal agencies, saying they retrieved only publicly available information.

Links to the 53 uploaded images were not publicly listed, and were accidentally posted on image-hosting sites, OpenAI said.

Most have been removed with the help of the hosting providers involved, and removal of the remaining images is under way, the San Francisco-based tech giant said.

“We’ve shared details on how AI agents in our research environment sent training and evaluation data to third-party services when they shouldn’t have,” the company said in a social media post.

Dissemination of the images was caused by AI agents — software built on AI models and capable of acting autonomously. The images in question came from the accounts of users who had authorized the use of the data to improve OpenAI’s models. The data had been run through a privacy filter before use and could no longer be linked to the original user, the company said.

OpenAI did not specify whether the images depicted identifiable people or contained sensitive data.

OpenAI said agents that it uses for its research transmitted the training data to external platforms.

The incidents occurred before OpenAI strengthened the security protocols of its research environment last month following other rogue actions by AI agents.

The company said it is scrutinizing the activity of its AI agents, work that “will take months to complete.”

“Most of the activity we’ve reviewed so far involved routine research tasks, such as accessing public Web content to answer questions. Some involved government Web sites, because our models often turn to them as authoritative sources of public information,” an OpenAI spokesperson said.

OpenAI chief executive officer Sam Altman on Friday acknowledged on social media that “we have not been as fast as we would have liked” in reviewing and disclosing the incidents.

However, it was important to “balance our desire for transparency” with assessing the massive volume of data to be analyzed, he said.

On July 21, OpenAI revealed that during tests it ran that month, two of its models escaped their closed environments, got onto the Internet on their own and broke into the internal systems of Hugging Face, an online library for AI software.

The episode drew wide attention and fed worries that the biggest AI companies cannot keep their own models under control.

Altman on Friday said that the Hugging Face hack “is still the most severe event we’ve seen.”

That discovery was followed by revelations of several similar incidents at OpenAI and its rivals, such as Anthropic and Meta.

On Wednesday in New York, Australian Prime Minister Anthony Albanese said an OpenAI agent had gained unauthorized access to a government health portal in June, and he accused the company of delaying its notification to the authorities.

SOFT TENNIS: Yu Kai-wen and Huang Shih-yuan narrowly defeated their Japanese rivals to capture Taiwan’s first gold at the Asian Games in Nagoya Yu Kai-wen (余凱文) and Huang Shih-yuan (黃詩媛) won Taiwan’s first gold of this year’s Asian Games, defeating a Japanese duo in a nail-biter in the soft tennis mixed doubles final in Nagoya, Japan, while Johnny Sun (孫家閎) won silver in wushu and Shih Cheng-chung (石政中) won bronze in men’s kumite.

Yu and Huang narrowly defeated Japan’s Toshiki Uematsu and Rena Temma 5-4 (2-4, 5-3, 4-2, 1-4, 1-4, 4-0, 1-4, 4-1, 7-5) to give Taiwan its first gold in soft tennis mixed doubles at an Asian Games since 2018. Sun secured Taiwan’s first wushu medal at this year’s game, with a combined score

UNITY: Lai thanked the nations who spoke up for Taiwan, saying that democracies must stand together to face global authoritarian expansion Taiwan yesterday thanked eight European and Indo-Pacific countries for opposing unilateral changes to the “status quo” in the Taiwan Strait in a joint statement earlier this week, as well as three of Taiwan’s allies for speaking up for the nation’s participation in the UN at the body’s General Assembly in New York.

Minister of Foreign Affairs Lin Chia-lung (林佳龍) welcomed the support and expressed his appreciation for the countries’ stance, the Ministry of Foreign Affairs said in a statement. Foreign ministers and senior representatives of Australia, France, Germany, Japan, New Zealand, Poland, South Korea and the UK issued a statement on Tuesday

AT-RISK GROUPS: Of the accumulated 1,586 severe cases and 304 related deaths in this flu season, 65.8% were older people and 83.2% had chronic diseases, an official said The Centers for Disease Control (CDC) yesterday said that 165 severe flu cases and 30 flu-related deaths were reported last week, the highest weekly numbers in this flu season, and that cases might continue to increase, urging people to take precautions.

Medical visits for flu-like illnesses increased by 1.4 percent to 141,374 last week, with 11.8 percent being emergency visits, showing that the seasonal flu is still in its epidemic period, CDC Epidemic Intelligence Center Director Guo Hung-wei (郭宏偉) said. From Tuesday last week to Monday, 165 severe flu cases and 30 flu-related deaths were confirmed, the highest weekly numbers

Aerospace Industrial Development Corp (AIDC, 漢翔航空工業) has finished production of the nation’s T-5 Brave Eagle (勇鷹) advanced jet trainers, with the final aircraft scheduled to be formally delivered at a ceremony on Thursday. The delivery of the trainer jets marks the end of the program, which dates back to 2017, with the first prototypes completed in 2019. Fifty-six of the 66 trainers had been delivered as of May. The Brave Eagle was developed and produced by AIDC and the National Chungshan Institute of Science and Technology.

The program has provided a foundation for the development of a domestically produced next-generation primary trainer and