While AI models are rapidly improving in their performance of complex tasks such as writing essays, generating art, or creating code, they have yet to master some simple skills that humans carry out with ease, researchers have found.

After investigating the limits of AI capabilities, a team from the University of Edinburgh has revealed that even state-of-the-art AI models are unable to reliably interpret clock-hand positions, or correctly answer questions about dates on calendars.

Unlike simply recognising shapes, understanding analogue clocks and calendars requires a combination of spatial awareness, context and basic maths – something that remains challenging for AI, according to the research team.

“Most people can tell the time and use calendars from an early age,” said Rohit Saxena, part of the research team from the University’s School of Informatics.

“Our findings highlight a significant gap in the ability of AI to carry out what are quite basic skills for people. 

“These shortfalls must be addressed if AI systems are to be successfully integrated into time-sensitive, real-world applications, such as scheduling, automation and assistive technologies.

To put current AI models to the test, the team examined whether systems that process text and images – known as multimodal large language models (MLLMs) – could answer time-related questions by looking at a picture of a clock or a calendar.

After testing various clock designs, including some with Roman numerals, with and without second hands, and different coloured dials, the researchers found that AI systems, at best, got clock-hand positions right less than a quarter of the time, with mistakes more common when clocks had Roman numerals or stylised clock hands.

According to the study, AI systems also did not perform any better when the second hand was removed, suggesting there are deep-seated issues with hand detection and angle interpretation.

The team found similar issues when asking AI models to answer a range of calendar-based questions, such as identifying holidays and working out past and future dates, with even the best-performing AI model getting date calculations wrong one-fifth of the time.

Recommended reading

• UK’s AI Research Ecosystem Needs Heightened Security, Report Finds
• Anthropic Reveals How AI is Really Being Used
• PwC Interview | How Agentic AI is Reshaping Business

Overcoming these limitations might enable future AI systems to power time-sensitive applications like scheduling assistants, autonomous robots and tools for people with visual impairments, researchers say.

“AI research today often emphasises complex reasoning tasks, but ironically, many systems still struggle when it comes to simpler, everyday tasks,” said Aryo Gema, a researcher from the University’s School of Informatics.

“Our findings suggest it’s high time we addressed these fundamental gaps. Otherwise, integrating AI into real-world, time-sensitive applications might remain stuck at the eleventh hour.”