Do you delete chat history after using ChatGPT? IT (information technology) experts say it’s a must to avoid potential data leak and security breach.
“The primary ChatGPT risks related to the security landscape data are privacy, confidentiality and cyber fraud risks. Any information entered into ChatGPT – if chat history is not disabled – may become a part of its training dataset.”
This warning was made by IT experts during the cyber security conference held as part of Intersec 2024, a three-day global trade fair in Dubai for safety and security protection that concludes on Thursday.
Stay up to date with the latest news. Follow KT on WhatsApp Channels.
Faheem Siddiqui, group director-Information Security at Majid Al Futtaim, said: “It is imperative for organisations to not only understand the security risks inherent in adopting Generative AI (artificial intelligence) but also proactively establish robust compliance frameworks.
“The convergence of technical innovation, ethical considerations, and regulatory alignment is pivotal in cultivating an accountable and transparent AI landscape in the Middle East,” he added.
Replicating human conversation
ChatGPT, a software developed by OpenAI to mimic human conversation, took the world by storm when it was launched in November 2022. Organisations, individuals and even students have embraced the AI-powered chatbot as a do-it-all virtual assistant to develop content, break down complex topics, and get insights from big data to produce analysis and forecast among other things.
ChatGPT works by predicting the best answer based on input data. This means, saving data is part of ChatGPT’s work. Experts note ChatGPT collects account information as well as your conversation history, as well as records such as email address, device, IP address and location, and any other public or private information that are used in ChatGPT prompts.
Collecting data is vital for ChatGPT to improve its ability in answering questions and helping users.
Averting security risk
Experts also argue ChatGPT saving data is not a security risk in itself but using the chatbot without implementing data security strategy is dangerous.
A simple step to keep sensitive data secure is to disable chat history. Security can also be extended by filtering what information go to the web by disabling sensitive information from being uploaded into ChatGPT.
At the Intersec 2024 cyber security conference, Grant Tuchten, portfolio director at Intersec organiser Messe Frankfurt Middle East, reaffirmed their “unwavering commitment to fortifying and safeguarding the UAE’s digital landscape”. He added 34 cybersecurity companies are exhibiting at Intersec 2024, fostering collaborations and sparking discussions on latest IT trends.
Dr Mohamed Al Kuwaiti, head of the UAE Cyber Security Council, for his part, noted: “The UAE Cyber Security Council is proud to support efforts to develop and further grow cyber awareness across global security sectors. Hosting the 3rd annual Cyber Security Conference, the UAE Cyber Security Council works closely with the show team to deliver new and innovative content areas, awareness sessions, and workshops, along with encouraging start-ups to flourish in cyberspace.”