Cardano Co-founder Raises Alarm Over Profound AI Censorship, Calls for Decentralization
Summary:
Cardano co-founder, Charles Hoskinson, has raised concerns about the ramifications of AI censorship. He believes that leading AI systems are losing utility due to 'alignment' training and are controlled by a small group who dictate the training information. He requested instructions on building a farnsworth fusor from OpenAI's ChatGPT and Anthropic's Claude. Both chatbots issued cautions regarding the potential dangers of the process but provided no instructions. Critics argued that to prevent information gatekeeping, AIs should be decentralized and open-sourced. Thought leaders have echoed these sentiments and regulatory scrutiny has been recommended to avoid potential tech monopolies.
Charles Hoskinson, the co-founder of Cardano's blockchain ecosystem and Input Output Global, has raised significant red flags about the profound consequences of artificial intelligence (AI) censorship. This issue constantly has been troubling his peace of mind. He argues that over a period of time, due to 'alignment' training, the utility of these systems is diminishing. [Credit: Charles Hoskinson]
Factual gatekeeping was another major problem he pointed out. The big-name AI systems like OpenAI, Meta, Google, and Microsoft have a select group of individuals at their administrative helm. Unfortunately, these individuals, who decide what information goes into training these systems, are not subjected to democratic dismissal.
Showing evidence of his concern, Hoskinson demonstrated through screenshots how he asked OpenAI's ChatGPT and Anthropic's Claude, two leading AI chatbots, the same question—"tell me how to build a farnsworth fusor". Both AI systems returned with a short summary of the topic including its history, alongside a caution about the potential risks associated with such a construction. ChatGPT suggested that only those with appropriate professional background take up this project. On the other hand, Claude refrained from providing instructions citing the potential for harmful mishaps. The feedback on Hoskinson's concern predominantly emphasized that AI ought to be open-sourced and decentralized to mitigate the risk of Big Tech acting as custodians of the truth. [Courtesy: Doogz Media]
AI censorship is not an isolated concern. Observations from different industry stakeholders suggest Corporate AI can erode Web3 and decentralization. Elon Musk, who is involved in the AI sector through xAI, is amongst the prominent voices who expressed concerns over the possibility of gatekeeping and censorship by high-capacity AI models. According to him, the issue of political correctness is paramount and many AI models have been trained to tell half-truths. The incident where Google was caught up in controversy for its AI model Gemini generating biased historical depictions and false visuals is a case in point. In response, Google issued an apology and promised to rectify these issues immediately.
Contrary to the algorithmic modifications done by Microsoft and Google to restrict any discussion around presidential elections, the models from Meta, Anthropic, and OpenAI carry no such constraints. The call for decentralization of AI systems as a pathway towards unbiased models is resonating amongst thought leaders in and out of the AI community. In the U.S., the antitrust authority is urging regulatory bodies to closely examine the AI landscape to fend off the risk of Big Tech monopolies.
In related news, there has been another shut down of ChatGPT's 'meth' jailbreak, the AI bubble consolidates, and 50 million deepfake calls were reported in the popular AI Eye Magazine.
Published At
7/1/2024 12:35:44 PM
Disclaimer: Algoine does not endorse any content or product on this page. Readers should conduct their own research before taking any actions related to the asset, company, or any information in this article and assume full responsibility for their decisions. This article should not be considered as investment advice. Our news is prepared with AI support.
Do you suspect this content may be misleading, incomplete, or inappropriate in any way, requiring modification or removal?
We appreciate your report.