OpenAI Introduces CriticGPT

Pivot 5: 5 stories. 5 minutes a day. 5 days a week.

1. OpenAI Introduces CriticGPT

OpenAI researchers have introduced CriticGPT, a tool that helps human trainers spot errors in AI models. CriticGPT produces thorough criticisms, particularly in code outputs, to improve the precision and dependability of AI systems. In experiments, human reviewers who examined ChatGPT's code outputs with CriticGPT performed 60% better than those without assistance.

The team has found that CriticGPT helps identify small errors in increasingly complex AI models, ensuring that sophisticated models stay in line with their intended behaviors and goals.

Read the full story here

2. New tool detects AI-generated videos with 93.7% accuracy

Columbia Engineering researchers have developed a new tool called DIVID to detect AI-generated videos. DIVID expands on Raidar, which detects AI-generated text by analyzing the text itself, without needing to access the inner workings of large language models.

The tool improves upon earlier methods that detect generative videos, which effectively identify videos generated by older AI models like generative adversarial networks (GAN). The new generation of generative AI video tools uses a diffusion model to create high-quality, lifelike videos, making it challenging to detect their authenticity.

Read the full story here

3. SoftBank forms AI healthcare JV in Japan with Tempus

SoftBank Group founder Masayoshi Son has announced a joint venture with Chicago-based health tech company Tempus to develop AI-powered personalized medical services in Japan. The company, known as SB Tempus, will focus on oncology, the largest cause of death in Japan.

Tempus, a genomic testing and data analysis company, will provide genomic testing, medical data aggregation and analysis, and AI insights for personalized treatments. SoftBank and Tempus hold a 50% stake, with SoftBank investing 30 billion yen ($188 million). The venture will start operations in August and offer three medical services to hospitals using AI to analyze personal medical data.

Read the full story here 

4. Cheap AI voice clones may wipe out jobs of 5,000 Australian actors

Voice actors are facing a potential threat from artificial intelligence, with their jobs being replaced by cheap generative AI clones. The Australian Association of Voice Actors claims that around 5,000 local voice actors are already in danger, with one national radio network investing in technology to replace human voice actors.

The group criticizes the development as a disappointing move from a player in an industry that has relied on voice artists for over 100 years. The association wants fair rules around AI technology use and protection for people's voices against misuse.

Read the full story here

5. MIT robotics pioneer Rodney Brooks thinks people are vastly overestimating generative AI

Rodney Brooks, a MIT Professor of Robotics Emeritus, believes that generative AI is not as capable as many suggest. He believes that humans tend to overestimate its capabilities and assign human capabilities to it. Brooks suggests that Robust.ai, a warehouse robotics system, uses massive data processing and AI optimization techniques to optimize for fast order completion.

He also emphasizes that robots should be integrated into solvable problems, such as cleaning up constrained spaces, rather than building a human-looking robot. The company designs robots for practical purposes related to warehouse operations.

Read the full story here

Advertise with Pivot 5 to reach influential minds & elevate your brand

Get your brand in front of 65,000+ businesses and professionals who rely on Pivot 5 for daily AI updates. Book future ad spots here.