Sam Altman of OpenAI Announces AI Agents Coming Next Year

Tech
By 24matins.uk,  published 4 October 2024 at 9h48, updated on 4 October 2024 at 9h48.
Tech

During the DevDay development event, Sam Altman of OpenAI announced the introduction of AI agents scheduled for release next year.

OpenAI: Pioneering the AI Revolution

OpenAI, a leader in artificial intelligence (AI), has announced plans to launch autonomous ‘agents’ by 2025. These advanced AI models can perform a variety of tasks independently and may soon be integrated into ChatGPT, a real-time AI chat application.

Autonomous Agents: Ushering in a New Era of AI

These agents are a significant step towards achieving General Artificial Intelligence (AGI), which could independently reason, plan, and act. We are currently in the second phase of AI, which involves reasoning before responding. The introduction of these agents will mark the transition to the third phase.

At the first OpenAI DevDay in San Francisco, CEO Sam Altman stated, “2025 will be the year when these agents become operational”, demonstrating this by showing how a voice assistant could autonomously make a phone call and order strawberries.

The Challenge of Aligning with Human Values

The deployment of these agents is not without challenges. The foremost among them is ensuring that agents align with human values and do not undertake harmful initiatives.

OpenAI emphasizes rigorous testing of its models before deployment, ensuring they meet safety criteria. This process has previously led to delays and the necessity for additional “safeguards” to prevent certain actions.

A Promising Future with AI Agents

Despite these challenges, the prospects of AI agents are promising. According to Sam Altman, “people will ask an agent to do something that would have taken them a month, and it will take an hour”. He envisions individuals having multiple agents performing various tasks, thus optimizing daily life.

In essence, the advent of these agents marks a significant stride towards truly autonomous and efficient AI, capable of performing complex tasks without human intervention, while also addressing ethical and safety challenges.