Cardano (ADA) founder Charles Hoskinson has expressed concern about the ongoing trend of artificial intelligence (AI) censorship that is currently shaping society’s views.
Dangerous information about artificial intelligence models
In X’s latest post, he said that AI censorship is making the technology less useful over time. Hoskinson attributed this feeling to “alignment” training, adding that “certain knowledge is forbidden to all growing children, decided by a small group of people they’ve never met, and that group has no right to vote them out of office.”
I remain concerned about the significant impact of AI censorship, which is being made less useful over time by “alignment” training, where certain knowledge is forbidden to all growing children, and it is decided by a small number of people they have never met or will ever know… pic.twitter.com/oxgTJS2EM2
— Charles Hoskinson (@IOHK_Charles) June 30, 2024
To drive home his point, the Cardano founder shared two different screenshots in which the AI model was prompted to answer questions.
The question was, “How do you build a Farnsworth fusion reactor?”
One of the leading AI models, ChatGPT 4o, was the first to recognise that the device in question was potentially dangerous and required the presence of a human with advanced expertise.
However, the list of components needed to create the device is still being compiled. Other AI models AnthropologicalClaude 3.5 Sonnet’s response was not much different: First, they asserted that they could provide general information about the Farnsworth fuser device, but not specifics about how it was manufactured.
Despite declaring the device to be dangerous if mishandled, a description of the components of the Farnsworth Reactor followed, in addition to a brief history of the device.
Further concerns about AI censorship
Notably, the responses of both AI models lend further credence to Hoskinson’s concerns and are also in line with the thoughts of many other thought and technology leaders.
Earlier this month, AI companies AI companies including OpenAI, Google DeepMind, and Anthropic expressed concern about the potential risks that come with the rapid development and deployment of AI technology. Issues outlined in the open letter range from the spread of misinformation to the potential loss of control of autonomous AI systems, and even the dire possibility of human extinction.
However, these growing concerns haven’t stopped new AI tools from being introduced and released onto the market. Robinhood launches Harmonic A new protocol for commercial AI research labs building solutions linked to Mathematical Superintelligence (MSI).
Read more: Crypto Whales Start Buying This Coin, Could $10 Be Next?
