#42 Story Points

Hi there! Are you on your fifth coffee break? Here’s Story Points with the news hotter and more energizing than espresso itself!

News sprint

1. How OpenAI's new agent works—a peek at the Operator's smooth (or not so smooth) functionality. #OpenAI
2. LinkedIn is facing a lawsuit for using users' data and direct messages to train AI models. #LinkedIn
3. According to the "State of Mobile 2025" report, users are more willing to spend extra on applications with AI features—the total spent in the category was over $1 billion in 2024. #AI
4. The latest news on the UnitedHealth data breach reveals that the cyberattack affected over 190 million Americans, doubling the previous estimates. #cybersecurity
5. DeepSeek, a Chinese startup developing open-source AI models, demonstrates its comparable performance to the other well-known chatbots, but at much lower costs. #Chatbots

Retrospective

Comment from LLI’s team about the unauthorized use of users data to train AI models:?

The unauthorized use of user data to train AI models is a pressing ethical, legal, and technical issue that threatens the integrity of the AI industry. At its core, it violates fundamental principles of data ownership, consent, and privacy, undermining trust between organizations and their users. Beyond the ethical implications, such actions expose companies to significant regulatory and reputational risks, particularly as global data protection laws become stricter.

From a technical perspective, using unauthorized data introduces vulnerabilities in AI systems. Models trained on improperly sourced datasets are often riddled with biases and inconsistencies that compromise reliability. Furthermore, these datasets can inadvertently include sensitive or proprietary information, leading to potential data breaches and liability for the companies deploying the model. Ensuring data is sourced transparently and ethically is critical for building robust, trustworthy AI systems that perform effectively across diverse use cases.