OpenAI has announced plans to develop an upgrade to its GPT-3 model which will allow its users to customize its responses. The move comes amid concerns about bias in artificial intelligence, and the company hopes that allowing greater diversity in views will help mitigate such concerns.
"This will mean allowing system outputs that other people (ourselves included) may strongly disagree with. Striking the right balance here will be challenging–taking customization to the extreme would risk enabling malicious uses of our technology and sycophantic AIs that mindlessly amplify people’s existing beliefs."
OpenAI also acknowledged that there will be limits to what the technology can still do. The company stated that it is impossible to entirely eliminate all biases and potential risks associated with AI technology, and so it will continue to put in place certain safeguards to minimize these risks.
In addition to allowing users to customize ChatGPT's behavior, OpenAI is also vested in improving the default behavior of their AI systems. The company is actively seeking public input on a range of topics related to its technology, including system behavior, disclosure mechanisms, and deployment policies.
We believe that many decisions about our defaults and hard bounds should be made collectively, and while practical implementation is a challenge, we aim to include as many perspectives as possible.
In the coming days, we can expect more bells and whistles to be added to ChatGPT and other integrated products like Bing's AI chat bot.
It's important to know that while AI can sometimes seem like an all-knowing oracle, at the end of the day, it's just good as the data it's trained on.
