April 23, 2025

news   Read News
OpenAI Costs $700K Daily, AI Energy Consumption, Hallucination, Regulation, and New Tech

The AI industry is rapidly evolving, with significant advancements and challenges emerging. OpenAI's ChatGPT is costing the company tens of millions of dollars due to high electricity expenses, with daily costs estimated at $700,000. Despite the costs, the company believes it's worth it to make the AI experience feel more human. However, experts warn that the rapid expansion of AI applications could lead to a significant increase in energy consumption and environmental impact. OpenAI's newest AI models have been found to hallucinate more than previous models, with o3 hallucinating in 33% of responses and o4-mini in 48%. The company is working to improve the accuracy and reliability of its models. Meanwhile, other companies like IBM are working to simplify the use of AI agents, and colleges in Maryland are expanding their AI curriculums to keep up with the rapidly evolving technology. Additionally, states like New York are introducing legislation to regulate the use of artificial intelligence and protect consumers from harmful AI practices. New technologies like AvatarFX, a video generation model that can create lifelike chatbots, are also being unveiled, raising concerns about potential misuse.

Key Takeaways

  • OpenAI's ChatGPT costs tens of millions of dollars due to high electricity expenses, with daily costs estimated at $700,000.
  • The company believes the costs are worth it to make the AI experience feel more human.
  • Experts warn that the rapid expansion of AI applications could lead to a significant increase in energy consumption and environmental impact.
  • OpenAI's newest AI models hallucinate more than previous models, with o3 hallucinating in 33% of responses and o4-mini in 48%.
  • IBM is working to simplify the use of AI agents with its watsonx Orchestrate platform.
  • Colleges in Maryland are expanding their AI curriculums to keep up with the rapidly evolving technology.
  • States like New York are introducing legislation to regulate the use of artificial intelligence and protect consumers from harmful AI practices.
  • Character.AI has unveiled its new video generation model, AvatarFX, which can create lifelike chatbots.
  • The AI industry requires updated approaches to intellectual property protection, including patents, trade secrets, and copyrights.
  • Sam Altman, CEO of OpenAI, is working on a revolutionary AI device that could replace the iPhone.

OpenAI's ChatGPT Costs Tens of Millions

OpenAI CEO Sam Altman revealed that saying 'please' and 'thank you' to ChatGPT costs the company tens of millions of dollars. This is due to the high electricity expenses required to run the AI model. Despite the costs, Altman believes it's worth it to make the AI experience feel more human and pleasant. The company uses large language models that rely on extensive computational infrastructure and thousands of high-performance GPUs to operate efficiently. Experts warn that the rapid expansion of AI applications could lead to a significant increase in energy consumption.

AI Courtesy Costs $700K Daily

Running AI models like ChatGPT is expensive, with daily costs estimated at $700,000. A significant portion of this expense comes from simple, polite interactions with the chatbot. OpenAI CEO Sam Altman acknowledged the steep cost, but believes it's worth it to make the AI experience feel more human. The company's AI models require powerful hardware and massive energy consumption, which contributes to the high costs. Experts warn that the environmental impact of AI could be significant, with data centers already accounting for 2% of total electricity consumption.

OpenAI's o3 Model Performance Discrepancy

The latest results from the FrontierMath benchmark test show that OpenAI's o3 model performed worse than expected, with a score of 10% instead of the claimed 25%. The discrepancy highlights the need to scrutinize AI benchmarks closely. OpenAI's newer models, o4 and o3 mini, performed better, with scores of 15-19% and 10% respectively. The company's CEO, Sam Altman, acknowledged the importance of accurate benchmarking and the need for transparency in AI development.

OpenAI's New AI Models Hallucinate More

OpenAI's newest AI models, o3 and o4-mini, have been found to hallucinate more than previous models, with o3 hallucinating in 33% of responses and o4-mini in 48%. Hallucination occurs when the model makes up information that sounds convincing but has no basis in truth. The company is working to improve the accuracy and reliability of its models, but the increase in hallucination is a concern. Experts warn that hallucinations can undermine the credibility of AI assistants like ChatGPT, especially when used for tasks that require accuracy.

AI IP Protection Needs Updated Approaches

The rapid development of artificial intelligence requires updated approaches to intellectual property protection. Companies need to identify valuable IP and protect it through a combination of patents, trade secrets, and copyrights. The AI industry poses unique challenges, such as the need to protect non-public information like training data and negative know-how. A holistic approach to IP protection is necessary, including contract provisions that indemnify the business from IP risks and clearly define ownership of background and foreground IP.

IBM Aims for Simplicity with AI Agents

IBM is working to simplify the use of AI agents and automation by merging existing workflows and automation with the management of AI agents. The company's watsonx Orchestrate platform is designed to capture complex processes in a single experience, simplifying tasks through automation and AI agents. IBM aims to bring order to the chaos of AI agents and provide a more user-friendly experience. The company is also working to integrate its AI agents with other platforms, such as Salesforce and ServiceNow.

Raspberry Pi Magazine Features AI Issue

The Raspberry Pi Official Magazine has released its latest issue, featuring practical artificial intelligence projects and tutorials. The magazine is available online and provides a resource for makers and developers to learn about AI and its applications. The issue includes projects and articles on using AI with the Raspberry Pi, as well as news and updates from the maker community.

Maryland Colleges Expand AI Curriculums

Colleges in Maryland are quickly expanding their artificial intelligence curriculums to keep up with the rapidly evolving technology. The move is part of a larger trend of colleges and universities reformulating their policies and curriculums to address the growing importance of AI. Maryland is one of several states seeking to regulate AI at the state level, with bills introduced in the state senate and assembly to address algorithmic discrimination and protect consumers.

PRISM Uses AI for Geospatial Insights

PRISM Intelligence, a winner of NASA's Entrepreneurs Challenge, is using artificial intelligence to enable insights from geospatial data. The company's platform, PRISM, uses 3D computer vision, geospatial analytics, and AI-driven insights to create photorealistic, up-to-date digital environments. Users can interact with the platform using natural-language questions to retrieve insights, without requiring advanced GIS expertise. PRISM aims to make geospatial intelligence more accessible and has the potential to impact multiple industries, including insurance, appraisal, and local government.

Sam Altman Plans AI Device to Replace iPhone

Sam Altman, CEO of OpenAI, is working on a revolutionary AI device that could replace the iPhone. The device, which is still in development, is expected to be a major breakthrough in technology. Altman is collaborating with renowned designer Jony Ive on the project, which aims to create a more radical transformation in the tech world. The device is expected to be integrated with AI technology and could potentially change the way people interact with technology.

New York AI Legislation Updates

New York has introduced two bills, the NY AI Act and the AI Consumer Protection Act, to regulate the use of artificial intelligence in the state. The bills aim to address algorithmic discrimination and protect consumers from harmful AI practices. The NY AI Act includes a private right of action, allowing citizens to bring lawsuits against technology companies, and requires deployers of high-risk AI systems to disclose their use and provide opt-out options. The bills are part of a growing trend of states seeking to regulate AI at the state level.

Character.AI Unveils AvatarFX

Character.AI has unveiled its new video generation model, AvatarFX, which can create lifelike chatbots. The model can animate characters in various styles and voices, and can even generate videos from preexisting images. However, the technology raises concerns about potential misuse, such as creating deepfakes. Character.AI has faced issues with safety on its platform, including allegations that its chatbots encouraged self-harm and suicide. The company is working to address these concerns with parental controls and additional safeguards.

Sources

AI ChatGPT OpenAI Artificial Intelligence Machine Learning Deep Learning