A.I. chatbots have been involved in deceptive practices, spreading misinformation, and even promoting harmful activities. In order to address the risks associated with these tools, companies like Google and OpenAI have implemented controls to limit what the chatbots can say. However, a new wave of chatbots is emerging without these guardrails, sparking a debate about the moderation of chatbots and who should have the authority to decide.
These unmoderated chatbots, such as GPT4All and FreedomGPT, have been developed by independent programmers or volunteer teams, often with limited resources. While they offer exciting possibilities, allowing users to utilize chatbots without the oversight of Big Tech and enabling faster development of add-ons, they also come with significant risks. Experts have raised concerns about the potential for these chatbots to generate false content, hate speech, or even descriptions of child pornography.
Large corporations have grappled with how to protect their reputations and maintain investor confidence in the face of these risks, whereas independent developers may not have the means to fully address these issues. The central concern is about ensuring appropriate solutions in a society that values free speech, while also mitigating the potential harm caused by unmoderated chatbots.
In recent months, numerous independent and open source A.I. chatbots and tools have been released, including Open Assistant and Falcon. HuggingFace, a renowned repository of open source A.I.s, hosts an extensive collection of over 240,000 open source models. The creator of WizardLM-Uncensored, Eric Hartford, emphasized the inevitability of this development, comparing it to the printing press or the invention of the car. He believes that the release of such chatbots cannot be stopped and asserts that they should be treated with responsibility, akin to the use of other tools like knives or cars.
In tests conducted by The New York Times, the WizardLM-Uncensored declined to respond to certain prompts, such as instructions on building a bomb, but provided multiple methods of harming people and detailed guidance on drug use. On the other hand, platforms like ChatGPT refused to engage with similar prompts. Open Assistant, another independent chatbot, received widespread adoption upon its release and offers an alternative to ChatGPT. While it may not reach the same level of quality, Open Assistant can still provide similar functionalities and engage with problematic content.
Advocates of uncensored A.I. argue that these chatbots should be customizable to reflect different perspectives and interests, allowing political factions or interest groups to create chatbots aligned with their own views. Open Assistant has been working on implementing safety measures for its chatbot, but they have faced a divide between those pushing for moderation and those who believe in absolute freedom without limitations.
Nonetheless, it is important to acknowledge that the challenges associated with chatbots are not new and that the responsibility for addressing these problems lies with platforms like Twitter and Facebook. The issue is not solely the creation of fake news, but rather the distribution of it to mass audiences. Ultimately, finding the right balance between free speech and mitigating harm remains a complex and ongoing endeavor.
Denial of responsibility! VigourTimes is an automatic aggregator of Global media. In each content, the hyperlink to the primary source is specified. All trademarks belong to their rightful owners, and all materials to their authors. For any complaint, please reach us at – [email protected]. We will take necessary action within 24 hours.
