Complaints about online platforms often go nowhere. AI could actually be part of the solution

Frustration is a common reaction when online platforms deny account access or refuse refunds. Over 1,780 complaints from Australians about issues with digital services were lodged with the telecoms industry ombudsman in 2023, a 20% increase from the previous year. Account access problems and billing disputes were the most frequent complaints, accounting for more than 70% of cases, and those from just five major tech companies—Meta, Google, Microsoft, Apple, and Hubbl—dominated the complaints.

However, current ombuds schemes lack the authority to make binding decisions on these issues, leaving many unresolved complaints and dissatisfaction rates as high as 83%.

Research shows that these unresolved digital issues cost the economy an estimated $497 million annually, with two-thirds of complainants leaving dissatisfied. In many cases, customers must navigate ineffective virtual customer service and AI chatbots that provide no real help or inaccurate information. The government’s scams prevention framework could serve as a model for resolving online platform complaints, but it only covers scams, not broader issues.

In Europe, platforms must provide proper complaint procedures and human oversight, as platforms must not rely solely on AI for decision-making.

To address this problem, AI chatbots and rule-based systems could be integrated into a dispute resolution scheme. These systems could help guide people through common problems by asking set questions and providing tailored information. However, generative AI’s ability to produce convincing yet incorrect legal advice poses a risk, particularly for consumers who lack legal expertise and are unaware when AI fabricates information.

To mitigate these risks, any integrated AI system must be overseen by human decision-makers, ensuring final decisions remain under human control.