NVDA 225.73 ▼2.01%GOOGL 338.36 ▼0.03%MSFT 493.95 ▼1.15%AMD 505.74 ▲5.90%INTC 104.47 ▲9.05%TSMC 439.00 ▲2.35%AMZN 256.97 ▼0.60%META 613.48 ▼0.53%AAPL 316.22 ▼1.17%PLTR 170.30 ▼2.31%
Markets at last close

OpenAI · Models

MIT study questions cognitive debt as OpenAI releases GPT-6 Astra

·1 min read

OpenAI released GPT-6 Astra, calling it the world’s most intelligent model and presenting it as a step into what President Greg Brockman called the “AGI era.” The model is described as capable of autonomously executing multi-step workflows across software, web pages and spreadsheets, with advanced agentic browser capabilities that let it interact with standard browsers to complete tasks.

The release has renewed attention on an MIT Media Lab study, “Your Brain on ChatGPT: Accumulation of cognitive debt when using an AI assistant for essay writing task.” Researchers used electroencephalography to monitor students writing essays with different levels of support, including an LLM, a search engine or no external tool.

The study found that using ChatGPT for writing tasks caused a 32% reduction in active mental effort and a significant drop in brain connectivity. Researchers said the technology does not make people “stupid,” “dumb,” or cause “brain rot,” but found that brain connectivity decreased as external support increased.

Participants who began with their own thinking and later used an LLM showed stronger recall and neural engagement than those who relied on the tool first. The findings point to a practical warning for advertising, copywriting and marketing: AI may be more effective as a refinement tool than as the starting point for creative strategy.

Originally reported by mediapost.comRead the source →
Related coverage
All OpenAI news →