A CHECKLIST FOR POLICY MAKERS ON PROTECTING CONSUMERS FROM THE RISKS OF AI
As consumers, we are likely to interact ever more frequently with artificial intelligence systems. In turn, these systems will undoubtedly get more powerful and extend to everything, from driving our cars to setting our insurance premiums. While people see a huge potential in this technology, they still have low trust in AI.
Trust in AI can only come about if people can rest assured that there are protections in place when something goes wrong. Here is a list of issues that any new regulation or law in Fiji must comprehensively cover, to address potential consumer issues. The recommendations are based on the European Consumer Protection Organization’s (BEUC) recommendations.
1. BROAD PRINCIPLES AND OBLIGATIONS MUST APPLY TO ALL AI
Any proposals for AI regulation must not only focus on ‘high-risk AI systems,’ it must also establish basic principles and obligations such as fairness, accountability and transparency to all AI. Laws should require that all AI-powered appliances or products, such as smart meters, connected toys, virtual assistants, or algorithms that organise what people see on social media, have these principles applied to them.
2. HARMFUL AI SYSTEMS: MORE OF THEM SHOULD BE BANNED
Lawmakers in Fiji must establish a list of AI practices that are banned, and this must be extended and strengthened to better capture consumer-related risks. Unacceptable practices such as social scoring, which is when an AI system evaluates the trustworthiness of an individual based on their social behaviour or their preferences, emotions, health or intelligence, should be banned when used by private as well as public bodies. Remote biometric identification systems such as facial recognition used by private entities in public places have no place in our society, as they are too intrusive and damaging of our fundamental rights. In addition to physical and psychological harm, the list of banned AI practices should include those that manipulate someone in a way that can cause them economic harm. Also, AI practices should be prohibited when they have the effect (instead of the intention) of causing physical, psychological or economic harm. AI which exploits any type of vulnerabilities, either temporary such as grief or emotional distress, or through digital asymmetry by using personalisation practices or persuasion profiles, must be forbidden. 3. CONSUMERS NEED STRONGER RIGHTS WHEN SUBJECT TO AI SYSTEMS Dealing with an AI system can be frustrating and problematic for consumers as they are not able to understand how the system functions or reaches its decision. In cases where an automated decision has a significant impact on consumers, they should have the right to be given a clear explanation about how an AI system affecting them works, and the right to object the decision. Any law enacted in Fiji must also grant consumers the means to seek justice and redress in case they are harmed. Consumers should have access to complaint mechanisms put in place by the provider of an AI system. They should also have a right to complain to a national authority, and to launch legal action, when an AI system or practice that affects them breaks the law. In this regard, consumer and civil society organisations should be able to represent consumers in exercising their rights or file complaints if an AI system breaks the law. Laws should also include a right to remedies, including receiving compensation for material or non- material damages suffered.
4. EFFECTIVE ENFORCEMENT NEEDED TO TACKLE BREACHES OF THE LAW
Any regulation will have to beef up enforcement by public authorities. One way to help national authorities is to create a highly specialised and independent body of designated technical experts. Their role would be to assist with the technical aspects of an investigation at national level, and have the competence to issue nonbinding opinions about specific cases brought up by the national authorities.
The Consumer Council advises that even though AI systems are still in their infancy and have yet to be extensively integrated into Fiji's consumer landscape, it's crucial to proactively establish regulations to anticipate and tackle potential challenges. By laying the groundwork now, we can ensure that consumers are protected and that the adoption of AI technologies proceeds ethically and responsibly. Moreover, such regulations can also foster innovation and competition in the AI industry, leading to better products and services for consumers in the long run. Therefore, it's imperative for policymakers and stakeholders to collaborate in crafting comprehensive frameworks that balance technological advancement with consumer protection.