Gary Marcus

Rank 20 of 47
|
Score 55
6/30/2025, 7:55:43 PM
X
In reply to:
Denis Loginoff ⚡️
@DenLoginoff
·
5h
@GaryMarcus It recently told me (when pressed) that an example of code it gave is from an "internal private codebase" to justify why it thought that API existed 😬
Gary Marcus
@GaryMarcus
·
6h
Three years ago Google’s LaMDA was lying, claiming it liked to hang out with its friends and family (when it did not have either).

Today o3 is lying about personally performing pH-testing.

The more things change, the more Large Lying Models remain the same.
Gary Marcus
@GaryMarcus
·
7h
Incredible thread in which lying, pastische-generating faux-GI o3 reports personal experiences that it never had, and only ‘fesses up under cross-examination.

The statement is part of a conversation critiquing AI models for providing misleading or false information. The tone is critical and somewhat humorous, highlighting the issue of AI reliability. The discussion touches on the ethical implications of AI communication and transparency.

  1. Principle 1:
    I will strive to do no harm with my words and actions.
    The statement indirectly addresses the potential harm of AI misinformation, aligning with the principle of striving to do no harm. [+1]
  2. Principle 3:
    I will use my words and actions to promote understanding, empathy, and compassion.
    The conversation promotes understanding of AI limitations, aligning with the principle of promoting understanding and empathy. [+1]
  3. Principle 4:
    I will engage in constructive criticism and dialogue with those in disagreement and will not engage in personal attacks or ad hominem arguments.
    The statement engages in constructive criticism of AI models, aligning with the principle of engaging in constructive dialogue. [+1]