OpenAI supports the idea of requiring advanced artificial intelligence systems to obtain licenses.

2023-07-24

OpenAI Drafts Internal Policy Memo Supporting Government Licensing for Advanced AI Systems OpenAI has drafted an internal policy memo indicating its support for the idea of requiring government licensing for anyone seeking to develop advanced artificial intelligence (AI) systems. The document also suggests that the company is willing to lift the veil on the data used to train image generators. Following a meeting between White House officials and tech executives, including OpenAI CEO Sam Altman, on May 4th, the creators of ChatGPT and DALL-E outlined a series of AI policy commitments in an internal document. The San Francisco-based company stated in the draft, "We commit to working with the US government and decision-makers worldwide to support licensing requirements for the most capable foundational models of future generations." The idea of a government licensing system developed by AI giants like OpenAI has raised concerns about potential conflicts with startups and open-source developers, who may perceive it as a means to make it harder for others to enter the field. This is not the first time OpenAI has proposed such an idea. During a May Senate hearing, Altman expressed support for the establishment of an agency that could issue licenses for AI products and confiscate them if anyone violated established rules. As Microsoft, Google's Alphabet Inc., and OpenAI are expected to publicly commit to safeguards for developing this technology in response to the White House's call, the release of this policy document coincides with the announcement. According to insiders, the two companies will pledge to develop and deploy AI responsibly. OpenAI cautioned that the ideas presented in the internal policy document differ from what the White House is about to announce and what other tech companies are proposing. Anna Makanju, OpenAI's Vice President for Global Affairs, stated in an interview that the company is not "advocating" for licenses but sees them as a "reality" for governments to track emerging systems. "For governments, it's important to recognize whether super-powerful systems that could have potentially harmful impacts are emerging," she said, adding, "If someone is not willing to self-report like we are, there are few ways to ensure that the government knows about these systems." Makanju clarified that OpenAI only supports a licensing regime for AI models more powerful than OpenAI's current GPT-4 and aims to ensure that smaller startups are not burdened with excessive regulation. "We don't want to stifle the ecosystem," she said. OpenAI also expressed its willingness in the internal policy document to be more open about the data used to train image generators like DALL-E and committed to incorporating source methods by the end of the year. Policymakers propose that transparent practices, where developers are accountable for their work and its sources, are crucial in preventing the spread of misinformation and bias in AI tools. The commitments outlined in the OpenAI memo closely align with some policy recommendations announced by Microsoft in May. OpenAI noted that despite receiving a $1 billion investment from Microsoft, it remains an independent company. The company revealed in the document that it is conducting an investigation into watermarking, a method to track the authenticity and copyright of AI-generated images, as well as detection and disclosure of AI-generated content. It plans to publish the results. OpenAI also stated in the document that it is open to external red teams—allowing people to enter and test vulnerabilities in its systems in various aspects, including offensive content, manipulation risks, and misinformation and bias. The company expressed support for establishing an information-sharing center for collaboration on cybersecurity. In the memo, OpenAI appears to acknowledge the potential risks AI systems pose to the job market and inequality. The company stated in the draft that it will conduct research and provide recommendations to policymakers to protect the economy from potential "disruption." Source: https://techxplore.com/news/2023-07-openai-idea-requiring-advanced-ai.html#google_vignette

RECENT AI NEWS

RECENT AI TOOLS