Do not rely on AI for betting, experts warn

Experts are warning the public against relying on artificial intelligence tools for betting as AI models claiming to predict football match outcomes with high accuracy gain traction on social media.

Sam Lam Sum, associate vice president of Hong Kong College of Technology, argued that while AI can analyze trends and behaviors on a macro level, accurately predicting single-match results is unreliable due to the unpredictable nature of sports.

"There are too many variables, such as player fitness, performance, weather, and tactics and it’s nearly impossible to predict the outcome until the final whistle," Lam told Sing Tao Daily, The Standard’s sister newspaper.

His warning came as advertisements for AI football prediction tools flood social media, claiming to use advanced data science, machine learning, and deep learning to provide betting tips, including match winners, point spreads, and goal totals.

One model, for instance, claims an 80 percent win rate, having predicted 785 matches in the past year.
It charges a minimum membership fee of HK$3,800 for 30 days but offers no free trials. 

A promoter contacted by Sing Tao Daily said, “football predictions are about long-term performance," advising users to focus on "long-term, stable returns" rather than quick riches.

Augusta Yim Ting-ling, a psychologist and service director at the Zion Social Service Yuk Lai Hing Counselling Centre, criticized the marketing tactics of these AI models, noting that they often attract novice bettors with their "scientific" packaging.

She questioned their credibility, citing variables like player morale and personal issues that can impact results.

"Can these platforms truly have such a comprehensive database? Building and maintaining AI models require substantial resources. Are user fees enough to sustain operations?" Yim asked.

She also warned that promoting AI tools as "professional investments" misleads users into believing they can achieve stable profits from betting. 

While her center has not yet received cases related to AI betting, Yim noted past scams involving "guaranteed win" schemes.

She urged increased public awareness, cautioning that simply telling people not to trust AI tips might backfire. Instead, she recommended providing more information to help the public discern their unreliability.

When queried about a specific football match, AI models like DeepSeek and GPT-4o mini declined to provide exact predictions, instead suggesting users consult professional analysis and prediction websites.

(Ayra Wang)