))
Why ChatGPT doesn't always tell the truth - Hallucinations, bias and lies
Manage episode 392714928 series 3415392
Contenuto fornito da David Nelson, CFA, David Nelson, and CFA. Tutti i contenuti dei podcast, inclusi episodi, grafica e descrizioni dei podcast, vengono caricati e forniti direttamente da David Nelson, CFA, David Nelson, and CFA o dal partner della piattaforma podcast. Se ritieni che qualcuno stia utilizzando la tua opera protetta da copyright senza la tua autorizzazione, puoi seguire la procedura descritta qui https://it.player.fm/legal.
In episode 2 of The Money Runner interview with Tech visionary Jeff Huber David and Jeff discuss the rising concern over biased or inaccurate output from some large language models.
Concerns Over AI Bias and Security Risks: The conversation highlights that 79% of senior IT leaders have concerns about potential security risks in AI technologies, and 73% worry about biased outcomes.
AI's Tendency to Confirm Existing Beliefs: An example is given where using AI to define a political outcome resulted in the AI seemingly confirming the user's pre-existing beliefs. This raises questions about AI's ability to provide objective information versus reinforcing subjective viewpoints.
AI Hallucinations and Data Dependency: The concept of 'hallucinations' in AI is discussed, explaining that AI systems can produce less reliable or factually incorrect results when dealing with data at the fringes of their training models.
The Challenge of Unbiased AI Systems: The discussion contemplates whether the goal should be to create completely unbiased AI systems or to accept some level of bias for potentially better outcomes.
00:00 Chat GPT lied
01:09 Hallucinations
02:35 Bias
05:22 How will small companies compete
07:50 Who is using A.I. and who's faking it?
09:37 A lot of money will be lost
11:54 So called A.I. experts
45 episodi
Condividi
