Tag: Accuracy

Mattel is the latest company to experiment with the generative AI tool ChatGPT to enhance their cybersecurity practices. However, concerns over accuracy and potential misuse have arisen. OpenAI reports that even successful models may produce false outcomes. Cyber teams are advised to consider specific questions with objective answers to reduce inaccuracies. Though some companies have banned the tool, Mattel's manager and others believe that human oversight is necessary before using ChatGPT or similar tools. Proper data restrictions are crucial. Follow along as we dive into generative AI!

CentML, a Toronto-based AI startup, has launched its platform to optimize machine learning models, making AI development more efficient. By enabling access to more efficient use of AI chips without affecting accuracy, CentML could potentially improve the broken marketplace for GPUs and reduce compute costs. With a powerful open-source compiler to automatically tune optimizations, CentML looks to capture market share from established competitors, unlocking more efficient compute for AI development. Help address worldwide GPU shortage now.

OpenAI's ChatGPT faces scrutiny from Dutch data protection authority over handling of personal data and potential offensive content. European regulators establish a joint task force to address concerns.

OpenAI, the company behind the AI chatbot ChatGPT, faces a libel lawsuit after providing false information about a legal case. The incident highlights the challenges involved in developing advanced AI models that balance creativity and accuracy. Researchers have identified the issue of hallucination, where AI models fabricate information, as a significant problem in the industry. OpenAI will focus on refining the ChatGPT model to reduce such incidents in the future.

The Dutch data protection authority has raised concerns regarding the handling of personal information by artificial intelligence companies, including creators of popular chatbot, ChatGPT. With 1.5 million users in the Netherlands, AP is asking for clarity on how the algorithm is trained, how data is collected and the appropriateness of the information provided. The watchdog also wants reassurance that sensitive data can be rectified or deleted. European privacy regulators have established a ChatGPT task force to address regulatory concerns. Users can now control their chat history to help train the system.