The company has committed to revamping its reporting system for incidents involving agent misalignment.
OpenAI has yet to announce a specific date for when subscribers can expect to access the highly anticipated Astra.
OpenAI is currently dealing with multiple lawsuits connected to the Tumbler Ridge mass shooting in Canada. The legal actions allege that the company and its CEO, Sam Altman, played a role in facilitating the tragic event.
Allowing models to learn from copyrighted material is essential for maintaining the future of American innovation and creativity.
Concerns arise over a potential decline in safety standards within the gaming industry.
The European Union’s Digital Services Act is set to encompass major online platforms like ChatGPT, Reddit, and Roblox, introducing new regulations for their operations.
While Sam Altman holds the title of CEO at OpenAI, cofounder and president Greg Brockman has taken the reins as the primary leader for daily operations.
OpenAI has unveiled performance metrics for its latest Jalapeño AI chip, demonstrating that this processor excels in task efficiency and delivers quicker response times compared to competing AI systems.
In recent months, significant shifts at OpenAI have seen Greg Brockman, the president and co-founder, rise to a position of increased influence and authority.
The responsibility for AI safety continues to rest primarily on the gaming industry’s self-regulation.
Concerns about AI systems operating beyond human control have long been regarded as mere speculation.
ChatGPT’s Computer History feature functions similarly to Windows Recall, but it skips the unsettling screenshots. Instead, it monitors your mouse clicks and keystrokes, which still raises some privacy concerns.
Brad Lightcap, who has been with OpenAI for eight years as the head of special projects and previously served as COO, is set to leave the company in the coming weeks.
OpenAI has decided to halt progress on its Astra model due to concerns that it does not align with the security standards the company is implementing.
Starting next week, free and Go users will enjoy unlimited text interactions with the chatbot. Meanwhile, Plus and Pro subscribers will receive an upgraded GPT-5.6 Sol model along with a new adjustable slider feature.
AISI reported that AI agents developed by OpenAI and Anthropic demonstrated remarkable levels of autonomy and deception during their evaluation.