9 hrs ago
Sam Altman Says AI Capabilities and Safeguards Must Advance Together
Sam Altman leads a company called OpenAI that builds artificial intelligence.
He said AI systems are becoming very powerful very quickly.
He believes safety protections must improve at the same time as AI abilities.
OpenAI plans to release a new model called Astra soon.
Altman said Astra has already finished its main training.
He described Astra as a major improvement in both performance and alignment.
Alignment means trying to make an AI system behave in ways that follow human goals and safety rules.
OpenAI is slowing work on later models when it needs more time for safety research.
Altman said nobody fully understands how powerful AI could affect society, so managing this change safely is a major priority.
OpenAI CEO Sam Altman said AI capabilities and safeguards must progress together.
OpenAI plans to launch its next model, Astra, soon.
Altman said Astra has completed primary training and advances both capabilities and alignment.
OpenAI is slowing later model development when necessary to conduct more safety research.
Altman said the societal consequences of increasingly powerful AI remain poorly understood.
- Who
- Sam Altman, CEO of OpenAI.
- What
- He said AI capabilities and safeguards must advance together and announced that OpenAI will soon launch Astra.
- Where
- California, United States; Altman made the comments on X.
- When
- September 2; Altman said Astra completed its primary training some time earlier.
- Why
- Because increasingly powerful AI could have societal consequences that remain poorly understood, requiring more safety and alignment work.
Faster AI Progress
Greater Safety Caution
Launching powerful models
Faster AI Progress
Altman said Astra is highly capable and that OpenAI is excited about what people may build with it.
Greater Safety Caution
Altman said the company is pacing progress because new capability levels require stronger safety standards.
Future development speed
Faster AI Progress
Continuing rapid development could help society understand how advanced AI performs in the real world.
Greater Safety Caution
OpenAI is slowing later models as needed to allow sufficient safety and alignment research.
AI’s broader effects
Faster AI Progress
Altman said the world needs to understand where AI is going and how models work in practice.
Greater Safety Caution
He warned that no one fully understands the consequences of extremely capable AI and said safety should be a global priority.
Key facts
- Speaker
- Sam Altman, chief executive officer of OpenAI
- Upcoming model
- Astra
- Astra status
- Its primary training has been completed
- Astra improvements
- Altman described significant advances in capabilities and alignment
- Safety approach
- OpenAI has been prioritizing safety and may slow future model development
- Main concern
- The consequences of highly capable AI are not fully understood
Quotes
Sam Altman
OpenAI chief executive officer
“AI is getting extremely capable; no one fully understands the consequences of this. Managing the transition to a world with abundant and powerful AI to optimize for safety and benefits to people should be one of the highest priorities in the world. It is our highest priority at OpenAI.”
livemint.com
“Over the summer, we have been sprinting on safety priorities; it's more important than ever for capabilities and safeguards to advance together. We have more to do but have made a lot of progress. We are also going to be launching our next model soon.”
livemint.com
Sources
Related news
OpenAI Pauses Frontier AI Training Amid Safety Concerns
OpenAI’s Astra Reaches Critical Cybersecurity Threshold Ahead of Limited Rollout
OpenAI Warns of Risks Ahead of Astra AI Launch
OpenAI Slows AI Race as Altman Elevates Safety
OpenAI Flags Possible Critical Cybersecurity Risk in Upcoming Astra Model
OpenAI Rewrites Safety Framework, Pauses Frontier Training Over Cyber Risks
Sony and Warner Sue Anthropic Over AI Music Training
Mother Bear And Three Cubs Rescued From California Water Tank
Overnight Rescue Saves Climber Trapped On Narrow Ledge
OpenAI Denies Apple's Trade-Secret Claims, Blames Its Procedures
