Artificial Intelligence (AI) is already part of consumers’ daily lives, yet the term encompasses current technology, that which is still under development, and that which belongs to science fiction although it is reasonable to think it may exist in our lives after a certain time.
AI’s advent and development brings incredible opportunities to consumers and companies and, as with any innovation, it also brings its share of risks and implications to consumers’ autonomy and self-determination, to their privacy, their ability to interact with products and services, and lastly, the ability to impute responsibilities to manufacturers if something goes wrong. Consumer organizations have a positive attitude towards change and innovation – DECO certainly does – and want to enthuse consumers about this change. Enthusing consumers requires guaranteeing some essential aspects and I believe we must act on at least two areas: Regulations and Knowledge.
In the last Websummit, which took place in Lisbon, one of the speakers compared AI’s current situation with the beginning of the industrial revolution. In those days, the lack of regulations clearly harmed companies and consumers and delayed the industry’s development, which is why I support the need to regulate activity in aid of consumers and professionals. The last economic and social crises unmistakably evinced how the lack or insufficiency of regulations seriously hindered not only consumers but also professionals and the development of the economic activity overall. I believe this to be a step of utmost priority so that consumers may trust the products and services offered to them. I also believe it wise to follow OCDE[1]’s recommendations, as they state AI regulation should obey to five principles based in values: 1. Inclusive growth, sustainable development, and well-being; 2. Human-centered values and fairness; 3. Transparency and explainability; 4. Robustness, security, and safety; 5. Accountability.
Although the latest developments offer some hope, the European Commission’s proposal on Artificial Intelligence (AIA) does not seem to follow the path set forth by OCDE, let alone the fair claims made by consumer organizations such as BEUC or DECO. For example, the proposal does not guarantee a high level of consumer protection, as consumers are in fact outright excluded from the scope of the Regulation’s application. The proposal also does not generally apply to AI’s several systems (only to those of high risk) and the methodology to list AI specific systems that the proposed regulation will apply to, will leave out all future AI applications that have not been developed yet. And while the proposal rightfully states a list of forbidden practices, the latter not only falls short from what is needed, but also from what already exists – for instance, there is no prediction for economic damage. Unlike the Commission’s proposal, the regulation in question must provide consumers with the needed rights and protections so they may safely use AI’s products and services. The regulation should also provide them with the adequate means to react to possible abuse and to enforce their rights[2].
“The eyes don’t see what the mind doesn’t know” (Mona Hanna-Attisha)[3]. More than any other technology, AI’s systems escape most consumers’ comprehension, and its consequences may be much more severe. This means that we must educate and train consumers on how AI works in products and services they purchase, for example in terms of their choices and information. Only then, can consumers understand the opportunities and risks stemming from these new systems. Only then, can they make conscious, and even free, choices.
Circling back to the comparison with the industrial revolution: the truth is nowadays, consumers are much more protected than back then, not only due to the existing legislation, but also due to the development of strong consumer organizations that watch over their rights and due to the existence of means of reaction, namely collective actions. When the Representative Actions Directive (RAD) is transposed, all European Union countries will at least have a collective means of enforcing consumer rights, which is a powerful way of enforcing consumer rights and ensuring that they are compensated for the damage they suffer. Consumer organizations tend to see this type of means as a last resort, but the absence of a robust and effective regulatory framework and the lack of transparency about AI systems can make it the only means available.
Luís Silveira Rodrigues, President of the Board.
DECO Associação Portuguesa para a Defesa do Consumidor ( Portuguese Consumer Protection Association)
Rua de Artilharia Um 4º, 1269-160 Lisboa – Portugal
Phone: +351213710212 / Email: lrodrigues@deco.pt
[1] OECD, Recommendation of the Council on Artificial Intelligence, OECD/LEGAL/0449 – https://oecd.ai/en/ai-principles
[2] For more in-depth information see BEUC – REGULATING AI TO PROTECT THE CONSUMER, Position Paper on the AI Act – https://www.beuc.eu/general/artificial-intelligence
[3] Mona Hanna-Attisha – “What the Eyes Don’t See: A Story of Crisis, Resistance, and Hope in an American City” – June 2018
