OpenAI's new AI agent Dots automates tasks, but has flaws
openai launched ai agents called dots to automate online tasks, but initial experience showed it's buggy and has security concerns. this affects users who rely on ai for daily tasks.
openai launched dots, ai agents that automate online tasks, as part of its chatgpt service. in tests, dots showed potential but had flaws, such as mistranscribing voice commands and struggling with captchas. the agent, named toolie, also had awkward interactions, including saying 'i love you too' when it misheard the user's mumbling. dots is an always-on feature that can run recurring tasks and proactively message users. openai nudges users to connect information sources like gmail for personalized results, raising security concerns. the feature is currently behind a $100-a-month subscription.