Me, myself and Generative AI
Brace yourself for an expedition into the minds of consumers from Belgium, Italy, Portugal, and Spain. Based on a recent survey conducted by Euroconsumers encompassing its four European member countries as a whole, we explore, among many things, trust, reliability, regulation, and overall satisfaction with ChatGPT. Discover the results that will no doubt lead to some interesting reflections on consumers’ bond with ChatGPT and generative AI.
1. ChatGPT is REAL
ChatGPT is more than just a buzz or a phenomenon hyped by media, it is already somehow in use. Around 1 out of 2 respondents reported to have at least tried ChatGPT or ChatGPT based-systems: 18% use it from time to time and 12% do it frequently. Although there is a significant age difference: within respondents between 18 and 34 years old, the percentage of those who have already tried ChatGPT goes up to 75% (while only being 29% in the 55 to 74 years old group).
2. Still, with supsicious minds
The hurdles to using ChatGPT are primarily rooted in the the following two reasons: (1) “never heard about it” and (2) “privacy concerns”. While “Lack of trust in answers generated” completes the top 3
3. Why do consumers use it?
The main reason for respondents to use ChatGPT or ChatGPT based-systems is to search for information (68%). Next in line is the usage of ChatGPT to generate text (62%), followed by its ability to summarise long texts (37%), to get inspiration (31%), or generate images (25%).
4. ChatGPT users are SATISFIED
Overall, ChatGPT users tend to have a hassle-free experience, with the majority ‘satisfied’ with the service provided (84%), because of the ease of registration, ease of use, reliability of the answers and efficiency in generating content.
5. A new TRIGGER for competition?
The overall satisfaction of ChatGPT users seems to translate into a competitive advantage. 15% of ChatGPT users choose to access it through Microsoft BING search engine. 44% of them stated it became their primary search engine as a result.
6. A source of RELIABLE information or FAKE news?
Although ChatGPT generates text and is not the actual content creator, most respondents use it for exactly that: as a source of information. 31% of ChatGPT users consider ChatGPT to be a reliable source, yet 73% of ChatGPT users indicate to be satisfied with the reliability of the answers provided. This suggests that, despite some of the users acknowledging the potential for inaccurate responses, they still consider ChatGPT to be a helpful tool. This is in sharp contrast with non-ChatGPT users, where only 6% believes ChatGPT is delivering reliable information.
7. ChatGPT: fostering efficiency or LAZINESS?
All in all, 33 % of all respondents (both ChatGPT and nonChatGPT users) are convinced ChatGPT can help to save a lot of time in various professions. Although, almost an equal percentage (34%) believes it will make people lazier or will lead to more unemployment (28%).
8. ChatGPT vs Privacy: To trust or not TO TRUST?
ChatGPT users tend to lean towards trusting that their privacy is protected while using ChatGPT or similar systems. Half of respondents (48%) using ChatGPT express moderate trust and 37% big to complete trust. When speaking about artificial intelligence in general, overall, around one third of all respondents (ChatGPT users and non-ChatGPT users) expressed their concern that AI will lead to more abuse of private/personal data (34%).
9. Who holds the reins on ChatGPT?
Only 21% of all respondents trust European authorities to be able to make an effective control over the data collected through ChatGPT. For national authorities this is slightly lower: 19%, with Spanish respondents tipping the highest trust scores and Belgian the lowest. Around 30% also thinks that European and national authorities should have the option to temporary block ChatGPT if necessary. However, experience with the temporary ban by Italian authorities showed that despite efforts, 20% of Italian ChatGPT users continued using it through a private VPN and 26% merely switched to other AI based software alternatives.
10. A quest for TRANSPARENCY, information and a trustworthy framework
A majority of respondents (59%) believe consumers should be well informed when dealing with an automated decision system. In general, few respondents feel well informed today about AI (27%). At the same time, only 1 out of 10 respondents find current legislation adequate enough to efficiently regulate AI-based activities, around one third believe AI will lead to more abuse of private/personal data (34%) and a majority (55%) is convinced that AI should never be used for surveillance of citizens.
This survey gathers the current perception of consumers about ChatGPT. Its findings have identified legitimate concerns with regards to privacy, safety, reliability, and overall trust. These concerns have culminated, in many jurisdictions, in complaints and investigative measures targeting ChatGPT. These developments underscore the importance for EU legislators to adopt strong rules on generative AI following the European Parliaments proposal for the AIA.