MIT Technology Review is reporting on several key developments in technology and health. It highlighted the career of Rupert Young, chief product officer of MaxMind, whose company uses IP-address location data and fraud prevention with its GeoIP tool, assisting streaming companies, security vendors, merchants and ad-network providers.
The publication also noted that large language models (LLMs) possess a fundamental flaw, making them vulnerable to attacks that can trick them into performing unauthorized actions, such as providing instructions for sabotaging an aircraft’s navigation system. Separately, Anthropic has developed a new technique to explore the internal workings of its LLM, Claude, uncovering a "hidden space" where the AI processes concepts.
In health news, a European fertility group is advocating for limits on sperm donations, citing instances where donor-conceived individuals have discovered hundreds of siblings. An international cap could help address this issue. Additionally, MIT Technology Review reported that artificial intelligence is more prone than humans to developing biases in hiring processes, not only learning existing stereotypes but also generating new ones.
Full Article: A new stamp on cyberfraud prevention