- Pascal's Chatbot Q&As
- Archive
- Page 7
Archive
GPT-4o: The use of AI for social media vetting, intelligence analysis, and ideological content filtering normalizes mass surveillance and digital profiling. AI-driven policies could lead to...
...an era of automated authoritarianism, where machine learning systems dictate who gets included or excluded in society, who is considered a security threat and which ideologies are allowed to thrive
One employee described GSAi as "about as good as an intern," delivering "generic and guessable answers." This suggests it may not yet be a robust tool, potentially leading to underwhelming results...
...or reliance on human oversight. Grok: "My take: it’s a cautiously positive development with significant caveats—promising in intent but shaky in execution so far."
GPT-4o: Wikipedia (Ex-)Founder Larry Sanger’s calls for Elon Musk to defend free speech and transparency in relation to Wikipedia editing are either highly naive or intentionally misleading.
At minimum, Sanger neglects or obscures the inherent contradictions of Musk’s complex relationships with government power and entities commonly associated with the Deep State.
Overall, X users in this context predominantly view Musk negatively regarding Nazi symbolism allegations, with a strong consensus among critics that he’s either intentionally signaling extremism...
...or irresponsibly flirting with it. The salute and AfD support are the most cited flashpoints, reinforced by far-right enthusiasm.
Asking AI: Trump, Musk, DOGE + Palantir, Anduril, Booz Hamilton = Deep State? Grok: The (MAGA) movement’s tolerance for such contradictions suggests broad support will persist.
Grok: If Trump and Musk spin DOGE’s NASA presence as a triumph over bureaucracy, most supporters will likely embrace it, downplaying Deep State connotations.
GPT-4o: This initiative is revolutionary for universities, impacting how students learn, how researchers make discoveries, and how institutions manage vast amounts of information.
OpenAI’s collaboration with universities is bridging gaps in knowledge access, healthcare innovation, and AI fluency for the future workforce.
Grok: The reuse of this ID across Musk’s personal projects and a federal government website (doge.gov) raises questions about the separation between his private ventures and public responsibilities.
Grok: This aligns with broader criticisms of Musk’s dual roles, as seen in posts on X about DOGE targeting agencies regulating his companies (e.g., FAA, SEC).
GPT-4o: Only 10% of companies have "very advanced" AI maturity. Given AI’s rapid adoption, a higher number was expected. Only 49% of companies measure AI ROI.
Larger companies ($1B+) are the least likely to have an AI council. While expected to be leaders, many still lack AI governance structures.
GPT-4o: If U.S. courts reject fair use defenses in AI training cases, there could be an incompatibility between U.S. and EU AI copyright laws​.
Even if licensing deals expire, AI models retain and benefit from extracted content, creating an unfair market dynamic​. The report argues that opt-out regulations flip copyright law on its head.
GPT-4o: If Microsoft is taking active steps to combat piracy of its own software using AI, there is a strong case for it to also: Train LLMs Exclusively on Licensed Data, Assist Rights Owners in...
...Fighting AI-Related Piracy, Apply Anti-Piracy AI to Content Owners' Needs. If it enforces licensing for its own products, it should set the same standard for AI training and outputs.
GPT-4o: The report highlights that major AI companies (e.g., Meta, OpenAI, Microsoft, Apple, and others) have used pirated content to train their generative AI models.
Sources of pirated content include classic pirate sites such as LibGen, Z-library, Anna’s Archive, Watchseries, OpenSubtitles, and shadow libraries.