OpenAI has unveiled GPT-6 Astra, a new artificial intelligence model designed to operate computers directly by interacting with user interfaces. The model moves the company beyond text and image generation, enabling the AI to navigate screens, click buttons, and type text to complete complex workflows.
The announcement comes amid an intensifying rivalry between OpenAI and its primary competitor, Anthropic, both of which are racing to perfect agentic AI—systems that can take independent action to achieve a goal rather than simply responding to prompts.
The ability of GPT-6 Astra to operate computers marks a shift from traditional Large Language Models (LLMs) to AI agents. While previous versions required specific API integrations or plugins to interact with other software, Astra can observe a computer screen visually and manipulate the cursor and keyboard as a human user would.
This development follows a trend of increasing autonomy in AI. Anthropic had previously introduced similar “computer use” capabilities, allowing its models to navigate desktops to perform research or fill out forms. OpenAI’s entry into this space suggests a broader industry push toward automating the actual execution of digital work, rather than just the planning of it.
Impact on Digital Workflows and Business Automation
For businesses in Nigeria and across Africa, the shift toward agentic AI presents both a productivity opportunity and a workforce challenge. The Nigerian tech ecosystem, which relies heavily on Business Process Outsourcing (BPO) and virtual assistance, may see a significant disruption in entry-level digital roles.
Tasks such as data entry, basic software testing, and administrative scheduling—roles commonly filled by remote workers in Lagos and Nairobi—can now be handled by AI agents that operate the same software tools used by humans. This creates a pressure point for the African freelance economy to migrate toward higher-value strategic roles that require human judgment and local context.
Conversely, for Nigerian entrepreneurs and SMEs, GPT-6 Astra could reduce the cost of scaling operations. Small businesses can now automate repetitive back-office tasks—such as updating inventory across multiple platforms or managing complex CRM entries—without the need for custom-coded software integrations.
The transition to agentic AI also raises critical security and safety concerns. Granting an AI model the ability to operate a computer implies a level of trust in the model’s ability to avoid errors and resist malicious prompts that could lead to unauthorized data deletion or financial transfers.
OpenAI has not yet detailed the specific safety guardrails implemented for Astra’s computer-use feature or the exact timeline for its general release to the public. The company is expected to roll out the feature in phases, starting with a limited set of enterprise partners to monitor for stability and security vulnerabilities.
Explore more Tech stories from Business Elites Africa.
