Session
Future & Sustainable Work in the World of Generative AI
Round Table - 90 Min
Our session’s aim is to demonstrate the positive uses of AI policy and technology from an institutional perspective. Last year, the organizers of this panel held a session that focused on dark patterns as an online consumer threat. This year we want to continue the conversation and discuss how AI, data science and other IT tools can identify, investigate and analyze online challenges facing consumers, from an enforcer perspective. To achieve a substantive debate, we have secured panelists from a range of key groups: academic, governmental, private and international organizations. Our speakers are leading experts in the topic, also able to cover a multitude of related aspects. We will conduct the session in three parts. First, we will have our panelists speak about their work, i.e. the OECD’s AI Policy Observatory and Network of Experts that looks at AI within public policy, the work of the EnfTech project hosted by the University of Reading (UK), the ARBUZ tool for detecting abusive clauses and the project on dark pattern run by UOKiK (Poland). This will provide an overview of the current situation. In the second part of the session, we will discuss how AI helps us understand and regulate AI techniques that cause risks to consumers, and find ways to monitor, understand and investigate them. We will invite audience members to actively engage in this part as well, to allow for a wider array of perspectives to be shared. The third part of the panel will be dedicated to discussing synergies and overlaps in the work of the panelists, identifying current key challenges and opportunities, and predicting future developments. We will conclude with a summary of the discussion, some suggestions of take away points for enforcers on both understanding AI risks for consumers and integrating AI into their daily work.
We plan to organize a Round Table Workshop in the hybrid format to facilitate the participation of both speakers and participants who would be present both online and onsite. We will put speakers in conversation with one another after a moderator introduces experts and their relation to the topic. We will then give opportunities for the audience to engage in the conversation through exchange of thoughts on both what was said, but also to introduce their own experiences and work. We are aware that the discussion has to take place with equal weight and equal opportunities. We will provide two moderators – one each onsite and online - who will jointly facilitate the discussion. One of them will “read the room” onsite, while the other one will encourage activity online, within the chat, in order to create a sense of representation to both groups of participants and to facilitate a hybrid discussion.
🔒Office of Competition and Consumer Protection
Martyna Derszniak-Noirjean, Director of International Cooperation Office, Office of Competition and Consumer Protection (UOKiK), Government, Eastern European States Piotr Adamczewski, Director of Bydgoszcz Branch Office, Office of Competition and Consumer Protection (UOKiK), Government, Eastern European States Ewa Sikorska, Office of Competition and Consumer Protection (UOKIK), Government, Eastern European States Aleksandra Mrozowska-Sroka, Office of Competition and Consumer Protection (UOKIK), Government, Eastern European States
• Piotr Adamczewski - Polish Office of Competition and Consumer Protection (UOKIK), Government, Eastern European States • Sally Foskett - Australian Competition and Consumer Commission (ACCC), Government • Christine Riefa – University of Reading, Academia • Angelo Grieco - European Commission, International Organization • Melanie MacNeil – Consumer Policy - OECD, International Organization • Kevin Luca Zandermann – Tony Blair Institute for Global Change, NGO
Piotr Adamczewski
Martyna Derszniak-Noirjean
Anna Bilska-Cicha
Targets: The main topic of our session is related to AI and dark patterns. Dark patterns in themselves can pose harm to internet users, specifically consumer, however AI, when used incorrectly, can also be detrimental and intentionally or otherwise, be bias towards a group of people. As one of the key focuses of our panel is discussing ethical use of AI to help combat these online practices, this will link to ensuring the elimination of discriminatory practices. In doing so, it will ensure a reduction of inequalities of outcome, linking to target 10.3.
Report
AI can also be a source of empowerment in consumer protection and an ally of consumer agencies if responsibly deployed. Consumer law policymakers and enforcers must adapt to the rapid development of AI tools in e-commerce and put safeguards in place for consumers.
In terms of AI, keeping the state of play is staying behind the technology and the new risks it carries for consumers. AI initiatives of consumer agencies should be encouraged and further developed, and so should the international collaboration of the enforcers and policymakers.
The round table session focused on presenting some of the AI initiatives and key areas of AI usage identified in relation to consumer protection. Speakers from international organizations (OECD, EC), national consumer protection authorities (ACCC, UOKiK), academia (University of Reading), and NGOs (Tony Blair Institute) discussed their current AI projects and expressed their opinions on the potential of AI in consumer protection.
The session was opened with a brief introduction to consumer protection, followed by a presentation of academic research and perspective on the use of AI in the Enforcement Technology (EnfTech) toolbox. The expert stressed that embracing the use of EnfTech is inevitable for enforcement agencies in order to keep digital markets in check. There are some discrepancies between the consumer agencies who start their technological reorganization on different levels, but the results of the EnfTech survey showed that even very little technology can do a lot to enforcement.
Next, the global and European perspective on AI tools was brought to the table. The speakers discussed AI impact on consumers, the e-tools developed by international organizations and the opportunities for consumer protection enforcers in the deployment of AI. Some of the e-tools mentioned in the presentation by OECD included the Polish system for intelligent contract terms analysis (ARBUZ), the fake reviews detector, Korea’s Consumer Injury Surveillance System as well as currently developed AI tools for empowering consumers to reduce energy consumption and make greener shopping choices. The OECD is also working on an AI Incidents Monitor for automatic global monitoring of AI hazards. The EC presented its eLaboratory and its main features that help collect evidence and conduct investigations. The eLab tools are available to CPC case handlers in the eLab environment. Another example of an intelligent tool used by the EC in the 2022 ‘Black Friday sweep’ is the price reduction tool that covers different websites, languages and currencies and presents a graphic representation of a price evolution and flags potential infringements. The EC also launched behavioural experiments with the use of AI to test the impact of online practices on consumers (e.g. labelling of advertising, dark patterns, use of cookies).
The session then focused on the presentation of specific tools and AI projects carried out by national consumer agencies from Poland and Australia. The Polish authority has recently implemented an AI-powered tool for the automatic analysis of terms of contracts and the detection of abusive clauses. The development of the ARBUZ system was co-financed by the EC. The tool assists casehandlers and will be further trained under supervision. Another EC-funded project taken up by the Polish consumer agency is focusing on dark patterns and the possible AI deployment to detect them. Polish experience with implementing AI into a consumer agency shall be concluded in a form of a white paper. From the Australian perspective, attention was drawn to streamlined webform processing and manipulative design detection. The techniques on this matter include entity extraction, classification and relevance. The expert also stressed the importance of data and giving more control to consumers over their data (Consumer Data Right).
A different ankle insight into AI in consumer protection was also delivered by an NGO. The recent Tony Blair Institute for Global Change report calls for a new policy agenda for the UK and the reorganization of institutions to deal with technological change. During the session, some of the governance challenges were discussed. The presentation also covered the adoption of computational antitrust by agencies in Finland and the UK using a supervised machine learning approach. The examples demonstrated that the cartels could have been caught using statistical methods. The presentation concluded with policy questions for AI deployment, including ethics of the algorithm, human oversight and LLMs.
The session was summed up with conclusions from the speakers on the future of AI in consumer protection. Businesses move quickly with the deployment of AI technology, but enforcement agencies have to step into this game towards a more proactive approach. The priority of consumer agencies should be collaboration, development of AI tools and increased use of technology in the course of investigations. The potential and the limitations of AI should also be thoroughly analyzed to make better use of the technology for consumer protection.