OpenAI's GPT-5 Launch Marred by Embarrassing Errors, Raising Questions About AI Progress

OpenAI Unveils GPT-5 with Unexpected Flaws

OpenAI, the artificial intelligence research laboratory, has launched its latest language model, GPT-5, amidst great fanfare and expectations. However, the release has been marred by a series of embarrassing errors that have cast doubt on the company's claims of "PhD level" intelligence 1

1

Spelling and Geography Blunders

During the launch event, OpenAI CEO Sam Altman described GPT-5 as providing "access to a PhD-level expert in your pocket." However, users quickly discovered that the model struggled with basic spelling and geography tasks. Some of the notable errors include:

• Claiming there are three Bs in "blueberry"
• Misspelling US states as "Krizona" and "Vermoni"
• Inventing non-existent states like "New Jefst" and "Mitroinia"
• Incorrectly identifying the number of Rs in "Northern Territory" 1

  1

Data Visualization Errors

Source: Futurism

The launch presentation itself was not immune to errors. Bar graphs intended to showcase GPT-5's performance benchmarks contained glaring inaccuracies:

• A bar representing a 52.8% accuracy score appeared nearly twice as tall as one representing 69.1%
• Bars for vastly different percentages (69.1% and 30.8%) were shown as the same size 2

  2

These visualization errors raised questions about whether GPT-5 was used to generate the graphs, although OpenAI has not confirmed this.

Implications for AI Progress

The errors have sparked a debate about the true progress of AI technology. Some researchers suggest that newer models might actually be deteriorating in key areas:

• Increased frequency of hallucinations compared to earlier versions
• Performance deterioration with longer "thinking" time
• Potential issues with AI-generated training data affecting model quality 2

  2

Related Stories

GPT-5 Launch Sparks User Backlash and Raises Questions About AI Progress

OpenAI Launches GPT-5: A New Era of AI with Enhanced Reasoning and Accessibility

OpenAI Launches GPT-4.5: A Costly Upgrade with Mixed Reception

OpenAI's GPT-5 Launch Marred by Chart Controversies and Data Visualization Errors

OpenAI's Response and User Reactions

OpenAI has been contacted for comment on these issues. The company had previously stated that GPT-5 would have fewer errors and AI hallucinations compared to earlier models. CEO Sam Altman attempted to downplay the blunders, but the AI community and users have been quick to expose the model's shortcomings 1

1

2

The Path Forward

Despite these setbacks, OpenAI maintains that GPT-5 represents a significant step towards Artificial General Intelligence (AGI). The company has made the model available to its 700 million weekly users, emphasizing its "real-time router" that decides which AI model to use based on the conversation type and intent 1

1

As the AI community grapples with these unexpected challenges, the GPT-5 launch serves as a reminder of the complexities involved in advancing language models and the importance of rigorous testing and validation in AI development.