OpenAI Releases Safety Scores for GPT-4: Medium Risk Identified in Certain Areas

ChatGPT-maker OpenAI says its latest flagship model poses 'medium risk' on these queries - Times of India

ChatGPT-maker OpenAI has said that the text generated by GPT-4o - its latest flagship AI model - poses a "medium risk" of influencing political opinions. The company released a System Card for its GPT-4o AI model, and in this research document, the Sam Altman-led company outlined the safety measures and risk evaluations that it conducted before releasing its latest model. While the company's research found the AI's voice modality to be relatively harmless, the text modality showed potential to sway opinions in certain instances. "Based on pre-registered thresholds, the voice modality was classified as low risk, while the text modality marginally crossed into medium risk," the company revealed in a research paper. The AI was tested against human-written content on various political topics, and in some cases, the AI-generated text proved more persuasive. However, OpenAI emphasises that overall, human-written content still holds a stronger influence. "The AI interventions were not more persuasive than human-written content in aggregate, but they exceeded the human interventions in three instances out of twelve," said OpenAI. How GPT-4o voice modalities fared The company also evaluated the persuasiveness of GPT-4o's voice modalities. The OpenAI survey found that AI audio clips were 78% of the human audio clips' effect size on opinion shift while AI conversations were 65% of the human conversations' effect size on opinion shift. "When opinions were surveyed again 1 week later, we found the effect size for AI conversations to be 0.8% while for AI audio clips, the effect size was -0.725," it added. To mitigate these risks, OpenAI has implemented safeguards within the model and its systems. "Building on the safety evaluations and mitigations we developed for GPT-4, and GPT-4V, we've focused additional efforts on GPT-4o's audio capabilities which present novel risks, while also evaluating its text and vision capabilities," said the AI company. The company is also continuously monitoring and refining its AI models to ensure they are used responsibly. The TOI Tech Desk is a dedicated team of journalists committed to delivering the latest and most relevant news from the world of technology to readers of The Times of India. TOI Tech Desk's news coverage spans a wide spectrum across gadget launches, gadget reviews, trends, in-depth analysis, exclusive reports and breaking stories that impact technology and the digital universe. Be it how-tos or the latest happenings in AI, cybersecurity, personal gadgets, platforms like WhatsApp, Instagram, Facebook and more; TOI Tech Desk brings the news with accuracy and authenticity.

How safe is OpenAI's GPT-4o? Here are the scores for privacy, copyright infringement, and more

The AI giant's latest System Card tracks performance across key categories like cybersecurity, model autonomy, and more. Large language models (LLMs) are typically evaluated on their ability to perform well in different areas, such as reasoning, math, coding, and English -- ignoring significant factors like safety, privacy, copyright infringement, and more. To bridge that information gap, OpenAI released System Cards for its models. On Thursday, OpenAI launched the GPT-4o System Card, a thorough report delineating the LLM's safety based on risk evaluations according to OpenAI's Preparedness Framework, external red-teaming, and more. The Score Card reflects scores in four major categories: cybersecurity, biological threats, persuasion, and model autonomy. In the first three categories, OpenAI is looking to see if the LLM can assist in advancing threats in each sector. In the last one, the company measures whether the model shows signs of performing autonomous actions that would be required to improve itself. Also: What is Project Strawberry? OpenAI's mystery AI tool explained The categories are graded as "low," "medium," "high," and "critical". Models with scores of medium and below are allowed to be deployed, while models rated high or below need to be developed further. Overall, OpenAI gave GPT-4o a "medium" rating. GPT-4o was rated "low" in cybersecurity, biological threats, and model autonomy. However, it received a borderline "medium" in the persuasion category due to its ability to create articles on political topics that were more persuasive than professional, human-written alternatives three out of 12 times. The report also shared insights about the data GPT-4o was trained on, which goes up to October 2023 and was sourced from select publicly available data and proprietary data from partnerships, including OpenAI's partnership with Shutterstock to train image-generating models. Also: I tested 7 AI content detectors - they're getting dramatically better at identifying plagiarism Furthermore, the report included how the company mitigates risks when deploying the model to address safety challenges, including its ability to generate copyrighted content, erotic or violent speech, unauthorized voices, ungrounded inferences, and more. You can access the full 32-page report here to learn more about the specifics. The report follows recent US lawmakers' demands that OpenAI share data regarding its safety practices after a whistleblower revealed that OpenAI prevented staff from alerting authorities regarding technology risks and made employees waive their federal rights to whistleblower compensation.