Sam Altman Reveals: OpenAI Holds Off on GPT-5 Training, Here's Why
OpenAI’s stance on the development of GPT-5, the next iteration of their language model, has been a subject of discussion among industry executives and academics, with concerns being raised about the rapid advancement of large language models.
In a conference held by the Economic Times, an Indian newspaper, OpenAI’s CEO, Sam Altman, addressed these concerns, stating that they have not yet commenced training for GPT-5.
Altman emphasized that there is still a significant amount of work to be done before they can embark on the new model, as they are currently focused on developing the necessary ideas.
Interestingly, in late March, an open letter signed by prominent figures, including Elon Musk and Steve Wozniak, called for a six-month pause in the training of AI systems that surpass the capabilities of GPT-4 in all AI labs.
Altman responded to this letter, acknowledging the concerns expressed but also highlighting the lack of technical nuance regarding the specific areas that necessitate a pause.
He clarified that OpenAI had not initiated training for GPT-5 and had no immediate plans to do so.
To address concerns raised by AI advocates, Altman stressed that OpenAI is actively taking steps to mitigate potential risks associated with AI.
These measures include external audits, red-teaming, and safety tests, which are being implemented alongside, rather than relying solely on, a pause in training.
Furthermore, Altman expressed OpenAI’s opposition to regulating smaller AI startups, instead advocating for self-regulation within the industry.
However, he emphasized the importance of applying regulations to larger organizations. Altman’s intention to engage with lawmakers and industry stakeholders globally is evident through his visit to India and other proactive efforts.
By engaging in these discussions, Altman aims to instil confidence in OpenAI’s commitment to collaboration with regulators.
During these meetings, Altman encourages policymakers to carefully consider the potential misuse and negative implications of widespread AI adoption, urging the implementation of appropriate guardrails to prevent unintended accidents.
In summary
While concerns have been raised regarding the advancement of large language models, OpenAI has not yet commenced training for GPT-5.
They are actively working on developing ideas and addressing potential risks through measures such as external audits and safety tests.
OpenAI’s CEO, Sam Altman, is engaging with global stakeholders to foster collaboration and advocate for responsible AI practices, encouraging policymakers to implement necessary safeguards in the face of widespread AI adoption.