ChatGPT Under Fire for Inaccuracy by EU Data Protection Board
The European Data Protection Board (EDPB) has raised concerns about inaccuracies and false information generated by ChatGPT, a chatbot developed by OpenAI. In a preliminary report, the EDPB task force investigating the chatbot's compliance with the General Data Protection Regulation (GDPR) stated that while transparency measures have been implemented to prevent misinterpretation of ChatGPT's output, they are deemed insufficient to comply with the data accuracy principle.
Background of the Investigation
The EDPB, composed of EU national data protection authorities and the European Data Protection Supervisor, plays a crucial role in ensuring data protection compliance across EU countries. Following Italy's temporary ban on ChatGPT in 2023, the EDPB launched a task force to collaborate and exchange information on enforcement actions against AI companies like OpenAI.
The preliminary report by the EDPB task force highlighted several key findings:
Key Findings of the Report
- Collection of Training Data: The report categorized different stages of data processing, including data collection, pre-processing actions, training, prompts, and ChatGPT output. It pointed out that the collection of training data through methods like web scraping poses risks to individuals' rights and freedoms.
- Lawfulness of Processing: The task force emphasized the need for safeguards to address the impact on data subjects, including defining criteria for data collection, filtering methods, and deletion of personal data collected via web scraping.
- Transparency and Fairness: The report highlighted the responsibility of OpenAI to inform users about the use of their data for training purposes and not transfer compliance burden to data subjects. It stressed the importance of compliance with GDPR principles, particularly in ensuring data accuracy and transparency in output creation mechanisms.
While OpenAI has presented measures to address these concerns, the effectiveness of these measures is yet to be evaluated by the EDPB task force.
The report also emphasized the need for compliance with GDPR's data accuracy principle, especially in providing proper information on the probabilistic nature of the system and potential biases in the generated text.
As the investigation by the EDPB task force continues, the focus remains on ensuring that AI technologies like ChatGPT adhere to data protection laws and uphold principles of accuracy, transparency, and fairness.
Support for Journalism
In 2016, The Securities and Exchange Board of India (SEBI) cautioned investors about online stock games involving the Bombay Stock Exchange and National Stock Exchange. The government also highlighted issues with caller ID practices related to fraudulent international calls using Indian mobile numbers.
MEDIANAMA is a leading source of information and analysis on Technology Policy in India.
