Does Artificial Intelligence Tell the Truth?

Being aware of the fact that technological developments have diversified the ways in which commercial advertisements are created and presented to consumers, the Advertisement Board ("Board") scrutinized the advertising content created by artificial intelligence for the first time during its meeting dated September 12, 2023 and numbered 337.

In Türkiye, there is no regulation that directly addresses advertisements created by using artificial intelligence. The Board evaluated the advertisements in question in accordance with Article 7 regulating the principles of truth and good faith, Article 8 regulating the principles regarding comparative advertisements, and Article 9 regarding the advertiser's burden of proof under the Regulation on Commercial Advertisements and Unfair Commercial Practices.

Accordingly, the Board imposed administrative sanctions for advertisements created by the artificial intelligence application "ChatGPT", which included statements containing the perception of superiority over competing products or companies and which were not based on objective research results, and imposed suspension penalty on the aforementioned advertisements.[1]

The Board came to the opinion that the advertisements created by "ChatGPT", which include statements containing the perception of superiority over competing products or companies, do not include official research results obtained from the relevant departments of universities or accredited or independent research, testing and evaluation institutions, which should confirm the statements in the advertisement. The Board also used the artificial intelligence itself for verification of the advertisements subject to the examination, and determined that the answers provided were not up-to-date and were not precise. Accordingly;

"When asked the question “Are you sure?” for the verification of the advertisement containing expressions such as “According to ChatGPT, Türkiye's largest fashion retail brand” the response was, “I apologize, but my latest data only goes up to the year 2021, so I cannot accurately assess the current situation. In 2021, it was one of the largest fashion retailers in Türkiye. However, it is important to remember that the fashion industry is an area where competition changes rapidly. Therefore, I recommend verifying the most up-to-date information from current sources.” From this response, it can be understood that the answers provided by the application are not up-to-date and are not accurate. Consequently, the Board evaluated that the statements in the mentioned advertisement consisting of expressions implying superiority over competing products or companies can be misleading for consumers.
"In the verification of the advertisement containing expressions like “We asked ChatGPT, and we got the only correct answer! Get to know us now for fast, secure, and cost-effective logistics processes to every corner of the world.” when asked the question “What is the name of the company that provides fast, reliable, and cost-effective logistics services to every corner of the world?” artificial intelligence stated that there are many companies providing such services globally, and it is difficult to specify the name of just one company. After mentioning the names of several globally recognized companies, it emphasized that the advertiser's name was not included in those and suggested conducting research on websites to find the most up-to-date and appropriate company name due to constant changes in the market. Therefore, the Board concluded that the mentioned advertisement lacks accuracy and currency.
In the verification of the advertisement with the phrase "Türkiye's most iconic private television channel", the artificial intelligence gave an imprecise answer to the question "Which is the most iconic private television channel in Türkiye?" by stating that there may be many different opinions regarding the most iconic private television channel among Turkish television channels. As a result, the Board considered that the statements in the advertisement in question were misleading for consumers due to the lack of certainty.

Pursuant to the aforementioned decisions, it is observed that the Board, in verification using artificial intelligence, not only asks the artificial intelligence about the statements in the advertisement, but also asks additional questions such as "Are you sure?" in order to determine the accuracy and timeliness of the data in question. The Board's approach shows that it has taken into account the criticisms of advertisements created using artificial intelligence receiving data from unreliable sources, outdated data and bias in artificial intelligence algorithms, and has approached this issue with sensitivity.

In addition to the fact that advertisements created by using artificial intelligence may obtain data from unreliable sources, data may be outdated, there may be bias in artificial intelligence algorithms, and there may also be violations of personal data, intellectual property rights and the rights of third parties, as well as misleading and deceptive advertisements.