British AI startup outperforms humans in forecasting competition

ManticAI, a British artificial intelligence startup co-founded by a former Google DeepMind researcher, has achieved eighth place in the international Metaculus Cup, outperforming numerous human forecasting professionals and enthusiasts.

Context and Background

The Metaculus Cup, organised by San Francisco-based forecasting company Metaculus, challenged participants to predict the likelihood of 60 diverse events throughout the summer. Questions ranged from political developments such as potential conflicts between Donald Trump and Elon Musk, to outcomes like Kemi Badenoch’s position in Conservative party leadership, and practical forecasts including wildfire acreage in the United States.

ManticAI’s system employs multiple AI agents working collaboratively to assess current situations, conduct historical research, model scenarios, and generate predictions. The startup breaks down complex forecasting problems into specialised tasks, then assigns them to different machine-learning models from providers including OpenAI, Google, and DeepSeek based on their particular strengths.

This achievement represents a significant milestone for AI forecasting capabilities, with some experts now predicting that artificial intelligence could match or exceed human forecasting performance by 2029, considerably earlier than previously anticipated.

Looking Forward

The competition results highlight AI’s persistent work capability as a key advantage, enabling systems to simultaneously tackle dozens of complex problems whilst revisiting them daily to incorporate new information. Ben Shindel, a professional forecaster who initially trailed behind AI systems before finishing ahead of ManticAI, noted the rapid progression: “We’ve really come a long way here compared with a year ago when the best bot was at something like rank 300.”

However, leading human forecasters maintain their edge in complex scenarios requiring judgment on sparse data. Warren Hatch of Good Judgment suggests the future lies in human-AI collaboration rather than replacement, stating: “The answer isn’t human or AI, but instead human and AI to get the best forecast possible as quickly as possible.”

Source Attribution:

Share this article