The Limitations ChatGPT Is Facing To Become The Most Brilliant AI – OsayWorld – The Light On a Hill

1984 is looming on us like dark clouds in a fierce storm. ChatGPT is making noise all across tech media sites seeking our attention with its smooth and plausible responses. Launched as a prototype in November 2022, I saw Elon Musk tweet about it and i knew it was a rocket waiting to be fired. The attention it garnered, however, is grounded in its potential to be the best chat bot and not its perfections. I’m covering some of the challenges the bot is seeking to overcome in order to hijack and rule humans.

Limitations of ChatGPT

  1. Nonsensical Answers
  2. Sensitive to Changes/Tweaks in Question phrases
  3. Verbose with Jargon
  4. The Responses are presumptuous
  5. Exhibit biased behavior
  1. Nonsensical Answers
    ChatGPT sometimes writes plausible-sounding but incorrect or nonsensical answers. Fixing this issue is challenging, as:
    (1) during RL training, there’s currently no source of truth;
    (2) training the model to be more cautious causes it to decline questions that it can answer correctly; and
    (3) supervised training misleads the model because the ideal answer depends on what the model knows, rather than what the human demonstrator knows.
  2. Sensitive to Changes/Tweaks in Question phrases
    ChatGPT is sensitive to tweaks to the input phrasing or attempting the same prompt multiple times. For example, given one phrasing of a question, the model can claim to not know the answer, but given a slight rephrase, can answer correctly.
  3. Verbose with Jargon
    The model is often excessively verbose and overuses certain phrases, such as restating that it’s a language model trained by OpenAI. These issues arise from biases in the training data (trainers prefer longer answers that look more comprehensive) and well-known over-optimization issues.
  4. The Responses are presumptuous
    Ideally, the model would ask clarifying questions when the user provided an ambiguous query. Instead, our current models usually guess what the user intended.
  5. Exhibit biased behavior
    While there have been efforts to make the model refuse inappropriate requests, it will sometimes respond to harmful instructions or exhibit biased behavior. The development team uses Moderation API to warn or block certain types of unsafe content, but they expect it to have some false negatives and positives for now. The company is collecting user feedback to aid in the ongoing work to improve the system.

ChatGPT (Chat Generative Pre-trained Transformer) is trained using human feedback and interactions. Since it has gone viral, given the public attention, investors are now on it, starting with Elon Musk whose interest seems to have brought ChatGPT into the limelight initially.

With all the funding soon to pour in, expect to see ChatGPT in your car, house, phone and even in your toilet seat reminding you to flash.

You are currently viewing The Limitations ChatGPT Is Facing To Become The Most Brilliant AI – OsayWorld – The Light On a Hill

Leave a Reply

Refund Reason

Select your currency
KES Kenyan shilling
×
×

Cart

Over 100 Services

Websites, Apps, flyers and More. Feel free to reach out and talk to us about your needs and goals; we’re here to help every step of the way.