The internet is becoming increasingly popular and a global protocol; however, this has also made accessing and using it more complex. Currently, there are billions of websites worldwide containing information, helping people connect, along with countless tasks on the internet that require dozens of steps to perform.
For the average person, mastering internet operations is simple, but it can become impossible for people with disabilities. Therefore, a group of scientists and engineers in Ohio (USA) are developing an online AI tool to simplify the process of using the digital world for everyone.
"For some people, especially those with disabilities, accessing the internet can be very difficult. We are becoming increasingly dependent on the digital world, both in our daily lives and at work, but this also brings more barriers and, to some extent, widens the digital divide," said Yu Su, a scientist and co-author of the study.
AI will solve the complexities of using the internet with just everyday language commands.
Yu Su and colleagues' project leverages the power of large-scale language models (LLMs) to create virtual employees capable of acting like humans when accessing the internet. The team demonstrated the model's ability to understand context and functionality, operating across various websites based solely on language processing and prediction.
Su stated that the project's success was largely due to the AI model's processing capabilities, accumulated through "learning" from the internet. The team had the AI perform over 2,000 tasks from 137 different real-world websites to fulfill training requirements. Some tasks were successfully completed, such as booking one-way and round-trip international flights, following celebrity accounts on X (formerly Twitter), and finding Netflix comedy films released between 1992 and 2017. Many tasks were considered difficult, such as booking an international flight, which required 14 steps.
"Training AI to perform tasks through language commands has only become possible thanks to the recent development of LLMs like ChatGPT," Su noted. Since OpenAI's chatbot was released to the public in November 2022, millions of users have used ChatGPT to generate automated content, from poetry and humor to cooking advice and disease symptom diagnoses.
In addition to assisting people with disabilities in using the internet, the AI developed by the team can also be used to enhance other artificial intelligence systems such as ChatGPT, transforming the internet into a tool with unprecedented power.
"We see great potential for improving performance, allowing us to focus on more creative aspects of our work. But this also comes with equally significant potential for harm," a representative from the development team stated. The risks mentioned include the misuse of users' financial information and the spread of misinformation.
Therefore, experts suggest that people should be cautious about these factors and make concerted efforts to try to minimize the risks. But Yu Su argues that human society will experience significant growth from the use of AI in commerce in the coming years, especially as the technology becomes more widely available to the public.
Source link
Comment (0)