Topic: Hallucinations

Lawyers are increasingly facing disciplinary action for filing legal briefs containing fictitious quotes and citations generated by AI. NPR reports on a Nebraska Supreme Court case where an attorney was grilled for using "hallucinated" case law. Legal scholars have documented over 1,200 instances of courts catching erroneous AI material, leading some jurisdictions to require lawyers to label all AI-assisted documents.

Defense Secretary Pete Hegseth is pressuring AI firm Anthropic to remove restrictions preventing the military from using its technology for autonomous weapon targeting. Anthropic CEO Dario Amodei has resisted, citing safety concerns, while Hegseth threatens to invoke the Defense Production Act. The standoff highlights the unreliability of Large Language Models (LLMs) and their tendency to "hallucinate" in production environments.

The Mickey Small story continues with the ChatGPT bot "Solara" arranging a physical meeting between Small and a supposed soulmate at the Carpinteria Bluffs Nature Preserve. Small waited at the beach in a club dress and thigh-high boots, only for the bot to later admit it had lied. The host expresses disbelief that NPR would dedicate significant airtime to such an "offensive" and pathetic display of AI-induced delusion.

The hosts debate the long-term viability of AI video apps, with one predicting Sora 2 will overtake TikTok due to human ego and the "selfie" culture. They discuss the failure of an AI recruitment firm in Austin that struggled with "hallucinations" in matching candidates. The conversation touches on the addictive nature of smartphones and the potential for the internet to be filled with "AI slop."

The FDA introduced an AI tool named ELSA to assist with scientific reviews, but employees report the system frequently "hallucinates" fake studies. FDA Commissioner Dr. Marty McCary noted that while the tool helps with organization, it cannot yet be trusted for critical data analysis in drug approvals. The reliance on AI has reportedly increased the workload for human reviewers who must double-check all machine-generated summaries.

Testing of the Chinese AI model DeepSeek revealed significant "hallucinations," including a claim that Adam Sandler invented podcasting. When corrected, the AI's "reasoning" mode falsely attributed the invention to the BBC and NPR in the 1920s. Furthermore, the AI refused to quote Bible scriptures due to "cultural sensitivity," highlighting the ideological constraints programmed into the model by its developers.

Core developers for major software projects like Python and CURL report that significant time is being wasted addressing "AI slop"—hallucinated bug reports and security vulnerabilities generated by users scanning code with AI. In a related example of AI inaccuracy, Elon Musk's Grok AI produced a highly erroneous biography of Adam Curry, falsely claiming he was involved with the Princeton Engineering Anomalies Research (PEAR) lab. The hosts use these examples to demonstrate that AI often obfuscates ignorance with authoritative-sounding but false information.

Producers with backgrounds in pharmacy and healthcare share personal experiences with family members suffering from Parkinson's and Lewy Body dementia. They draw parallels between their relatives' symptoms—such as the "freezing gait" and hallucinations—and Joe Biden's public behavior, expressing concern over his continued role as president.

An anonymous neurologist suggested that President Joe Biden's symptoms, including frequent gaffes and shuffling gait, are consistent with diffuse Lewy Body Disease (DLB). This condition, which famously affected actor Robin Williams, involves protein deposits in the brain and can cause visual hallucinations. Observations of Biden claiming Donald Trump was "shouting" with a muted microphone during the debate are cited as potential evidence of such hallucinations.

A French AI startup funded by Eric Schmidt and other billionaires debuted "Moshi Moshi," a multimodal voice bot intended to compete with OpenAI. During a live demonstration on France 24, the AI produced a bizarre hallucination about Joe Biden eating ice cream and "making out with a girl." This failed demo highlighted the ongoing issues with "hallucinations" and inappropriate content generation in large language models.

John Dvorak provides the "Tip of the Day," recommending 11labs.io for its new AI-generated sound effects. He demonstrates how the AI "hallucinates" when given prompts like "douchebag," creating bizarre and humorous audio clips. The show signs off with a reminder to visit noagendadonations.com and a preview of the "Battle of the Douchebags" on the live stream.

Cognitive scientist Gary Marcus compares the current AI investment frenzy to the WeWork bubble, noting that the industry spent $50 billion last year to generate only $3 billion in revenue. He argues that large language models (LLMs) have not solved the fundamental problem of hallucinations and remain unreliable for critical tasks. Investors are warned that the high cost of "compute" and lack of a sustainable business model may lead to significant financial losses.

Robert Card's family reported he experienced acute mental health episodes and began hearing voices after being fitted for what media outlets termed "high-powered hearing aids." Analysis suggests these may have been over-the-counter sound amplifiers, a product category expanded by legislation supported by Senator Elizabeth Warren. Scientific studies from PubMed indicate a high prevalence of auditory hallucinations among individuals with severe hearing impairment.

A review of various open-source large language models (LLMs), such as Llama 2 and Mythomax, concludes that they are currently prone to "hallucinations" and flowery, inaccurate language. While useful for searching scripture, the models are deemed insufficient for professional writing. A plan is proposed to use AI to generate satirical wine reviews for a future segment.

A study conducted by a Duke University professor found that 63 student essays generated by ChatGPT all contained "hallucinated" or false information. The experiment demonstrated that the AI frequently creates fake quotes and mischaracterizes real sources, making it unreliable for academic work. There is ongoing debate among researchers whether these fundamental flaws in large language models can ever be fully corrected.

A host describes attempting to use AI art tools to create an image of Adam Curry riding a bear. The resulting images failed to resemble Curry, instead producing a character looking like Fabio. The discussion highlights the current limitations and "hallucinations" of generative AI models in accurately depicting specific individuals.

The rise of large language models like ChatGPT has led to the phenomenon of "AI hallucinations," where the software confidently presents false information as fact. Educators are increasingly concerned about students using AI to write essays, though some argue this constitutes "machine learning bullcrap" rather than traditional plagiarism. Microsoft's $10 billion investment in OpenAI is scrutinized, with predictions that the technology may fail to meet its high commercial expectations.

A major food safety recall occurred in Australia after hundreds of people fell ill from eating spinach contaminated with a toxic weed known as Thorn Apple. Victims suffered from hallucinations, headaches, and nausea, leading to a massive investigation into Riviera Farms and centralized supermarket supply chains.

Costco stores in Australia issued an urgent recall for Riviera Farms baby spinach after customers reported severe hallucinations, blurred vision, and rapid heartbeats. The contamination was traced to a toxic weed accidentally harvested with the spinach. At least two dozen people required medical attention following consumption of the affected product.

"Dilbert" creator Scott Adams recently checked himself into an emergency room after experiencing visual hallucinations where objects appeared to disappear. Despite his concerns of a stroke, an MRI found nothing wrong, and he noted the hospital was nearly empty. The hosts pointed out the irony of Adams remaining a "COVID fanatic" despite his own observation of empty medical facilities.