Berlin When dealing with artificial intelligence (AI), Germany’s top consumer advocate Ramona Pop has spoken out in favor of comprehensive rules. AI systems like ChatGPT have the potential to “make everyday life easier and improve it,” Pop told Handelsblatt. But “binding rules are needed so that technical development doesn’t gallop away from us”. It is therefore right, according to Pop, that the EU is now moving forward.
Digital Minister Volker Wissing (FDP) hopes for “opportunity-oriented rules”. Especially with a view to small and medium-sized companies and start-ups, you have to “find a good balance between transparency obligations and freedom,” Wissing told the “Tagesspiegel”. The aim of the law is to help “so that AI applications are increasingly being developed in Europe and Germany in the future”.
Consumer advocate Pop assumes that society will change “radically” through AI. In everyday consumption, for example, there is “the risk of manipulation, fraud or discrimination by algorithms,” she said.
Exactly what the risks are – an overview:
Image recognition systems with artificial intelligence
AI-controlled face and emotion recognition can be problematic, for example in the finance and insurance sectors. “Such models can be used to find out whether I’m lying or not, for example if I’m supposed to describe damage to my car,” explained consumer advocate Pop. The problem is that applications for emotion analysis are “often immature”. “Data is misinterpreted and then leads to assumptions about an individual that are inaccurate and can lead to incorrect results.”
>> Read also: Why AI companies are demanding stricter laws and politics aren’t delivering
Product reviews by AI bots
Consumers hope that online product reviews will provide useful information for their own purchasing decisions. However, these reviews can also be purchased or fake. Pop speaks of “systematic risks” for consumers if AI bots are used on a larger scale.
Then legal questions would become even more relevant, said the consumer advocate. “Who is actually liable for false or even fake product reviews that suggest something that does not correspond to reality?” Questions like these would have to be clarified within the framework of regulation.
Online pricing with the help of AI
The annoyance for consumers begins when the prices for the same product differ significantly. One then also speaks of “dynamic pricing”. The decisive factor here is an AI-based analysis of factors such as demand, supply or competition. “That can be manipulative or even fraudulent,” states consumer advocate Pop.
In Pop’s view, what is and isn’t allowed should be clearly defined so that the consumer isn’t left behind. “The best thing would be to be as transparent as possible when it comes to pricing, so users understand how AI affects prices,” she said.
Price Comparison Sites
Algorithms are also used here. The consumer advocates have already taken legal action against comparison portals if, from their point of view, no objective price comparisons have taken place. You have to be able to rely on the reliability of the portals, Pop emphasized. That’s why transparency is the order of the day here, too.
“It must be understandable how such platforms come up with their recommendations and what influence, for example, commission payments from the product providers have on the displayed ranking,” said the VZBV boss.
Personalized advertising
Such advertising, which can already be created today with the help of cookies, will be even more targeted in the future with AI. “The risk of manipulation will therefore increase,” believes Pop, for example if the system determines based on the search history that a person may be particularly vulnerable in certain situations.
This can happen, for example, when the person is looking for medication for memory disorders. Then she could suddenly get offers on similar health issues, Pop explained.
How can consumers be protected from AI risks?
“Strong consumer rights” are necessary towards operators of AI systems, said Pop. “There must be a right to correction and deletion so that those affected can defend themselves against AI-generated false reports or manipulation.”
The topic of AI must also be “included in the annex to the European directive for class actions,” demanded the consumer advocate. “Then we have the opportunity to bring class action lawsuits against certain providers on behalf of injured consumers and can claim compensation.”
>> Read also: If the AI becomes a fraud – you should know these tricks
The head of the Agency for Leap Innovations (SprinD), Rafael Laguna de la Vera, spoke out against rules leading to a technology ban. “That makes us more stupid and therefore more defenseless,” he told the Handelsblatt. He suggested that AI results can be evaluated according to their output and based on the question of whether they follow our values and laws.
In this context, Pop advocated the establishment of an AI supervisory authority. “We need to have a discussion about how AI rules, risks and problems should be evaluated and what conclusions should be drawn from them,” she said. “To monitor compliance with rules, a European authority might be useful in addition to national supervision.” That must be decided quickly. Because this is also important for the social acceptance of this new technology.
More: The Germans’ fear of artificial intelligence