OpenAI تكشف عن LifeVal، اختبار يقيس قد إيش الذكاء الاصطناعي يقدر يساعدك في أمورك اليومية، مثل الميزانية والمشتريات والخدمات الحكوميةOpenAI unveils LifeVal, an evaluation of how well AI can help with everyday life, such as budgeting, shopping and government services

OpenAI has published details of LifeVal, an evaluation it built to measure how well its models can help people with everyday tasks, such as finding childcare, accessing government services, creating a budget, modifying recipes, ordering groceries and designing workout plans. The company says it used it over the past year to track and improve its models' performance on these tasks.

To choose the tasks, the company drew on two large studies of how people spend their time: the American Time Use Survey and the Multinational Time Use Study. It then ran a new survey of 3,995 people to find where they most wanted help. The tasks cover areas such as caregiving, health, education, household administration, personal finances, shopping and government services.

The company shared example tasks, each with its own score, such as enrolling in the right kindergarten after moving (93.1%), organising volunteers to help neighbours before a flood (98%), finding a home in Luzerne County in the US that fits a budget (100%), and a care and meal plan for a parent newly diagnosed with diabetes (90.4%). The page gives no overall score for the evaluation and does not name the model tested, only saying "our current models".

OpenAI says that over the past year it trained its models on browser use, computer use and connectors, so ChatGPT can carry out tasks itself rather than just give advice. Note that all the results come from the company itself and have not been independently verified, and the published examples are based on life in the US, so we don't yet know how it would do on tasks here, such as our own government services.