ChatGPT: Everything you need to know about the AI-powered chatbot

When an LLM doesn’t find the exact best answer for your question, it’ll give you one that sounds correct but isn’t. Or it could make something up out of thin air that appears plausible. Some of these errors, called hallucinations, are pretty obvious.

Before generative AI, the key was to judge the source, Mahadevan said. If you see something on Facebook or in a Google search result that claimed a celebrity or politician has died, you could check the background of the source providing the information to see if it’s reliable. The only source is your cousin’s neighbor’s ex-husband? Another important step is ramping up scarcity — that means taking away resources so employees need to do more with less. At the same time, it’s essential to redesign work processes hand-in-hand with those employees who are on the front lines. Don’t just identify laborious processes where manual work can be automated — identify the parts where humans bring value.

• OpenAI has effectively canceled the release of o3 in favor of what CEO Sam Altman is calling a “simplified” product offering.
• Some ChatGPT users have noticed a new feature called “Study Together” appearing in their list of available tools.
• Our readers wanted to understand the broader implications of this specific issue, and asked Smart Answers to source wisdom from our decades of reporting on IT M&A.

OpenAI announced a new AI “agent” called deep research that’s designed to help people conduct in-depth, complex research using ChatGPT. OpenAI says the “agent” is intended for instances where you don’t just want a quick answer or summary, but instead need to assiduously consider information from multiple websites and other sources. The Arc Prize Foundation, which develops the AI benchmark tool ARC-AGI, has updated the estimated computing costs for OpenAI’s o3 “reasoning” model managed by ARC-AGI. The organization originally estimated that the best-performing configuration of o3 it tested, o3 high, would cost approximately $3,000 to address a single problem. The Foundation now thinks the cost could be much higher, possibly around $30,000 per task. Many chatbots and other gen AI models today will provide citations in their responses to you, although you might need to ask for them or turn on a setting.

It offered a deep, detailed analysis of the vehicles. What it didn’t do was recommend random other vehicles it found on the internet — which Pearce wanted to avoid. « When you’re prompting it, adding all these caveats is really important, » she said. You can also ask it to play a certain role to get answers in the right tone or from the perspective you’re seeking.

OpenAI promises to make changes to prevent future ChatGPT sycophancy

Aidan Clark, OpenAI’s VP of research, is spearheading the development of the open model, which is in the very early stages, sources familiar with the situation told TechCrunch. OpenAI has announced its plan to make changes to its procedures for updating the AI models that power ChatGPT, following an update that caused the platform to become overly sycophantic for many users. OpenAI has started using Google’s AI chips to power ChatGPT and other products, as reported by Reuters. The ChatGPT maker is one of the biggest buyers of Nvidia’s GPUs, using the AI chips to train models, and this is the first time that OpenAI is using non-Nvidia chips in an important way.

Study warns of major risks with AI therapy chatbots

AI can also surpass human scale, even open up new applications that were not previously possible. For instance, makeup company L’Oreal launched a virtual beauty advisor to scale that exclusive service beyond their retail locations, reinventing the way they think about interacting with their customers at scale. I began as a cognitive psychologist studying language—the science underpinning today’s large language models (LLMs)—and later led teams at YouTube, Spotify, and Instagram.

Support for software support

OpenAI rolled out new tools designed to help developers and businesses build AI agents — automated systems that can independently accomplish tasks — using the company’s own AI models and frameworks. The Responses API effectively replaces OpenAI’s Assistants API, which the company plans to discontinue in the first half of 2026. OpenAI has updated its AI voice assistant with improved chatting capabilities, according to a video posted on Monday (March 24) to the company’s official media channels.

Epoch AI, the research institute behind FrontierMath, discovered that o3 achieved a score of approximately 10%, which was significantly lower than OpenAI’s top-reported score. After introducing a data residency program in Europe in February, OpenAI has now launched a similar program in Asian countries including India, Japan, Singapore, and South Korea. The new program will be accessible to users of ChatGPT Enterprise, ChatGPT Edu, and API.

“We are getting things under control, but you should expect new releases from OpenAI to be delayed, stuff to break, and for service to sometimes be slow as we deal with capacity challenges,” he wrote. OpenAI says its o1-pro uses more computing than its o1 “reasoning” AI model to deliver “consistently better responses.” It’s only accessible to select developers who have spent at least $5 on OpenAI API services. OpenAI charges $150 for every million tokens (about 750,000 words) input into the model and $600 for every million tokens the model produces. It costs twice as much as OpenAI’s GPT-4.5 for input and 10 times the price of regular o1. Lightcap, who previously worked with Altman at Y Combinator, joined the Microsoft-backed startup in 2018. OpenAI also said Mark Chen would step into the expanded role of chief research officer, and Julia Villagra will take on the role of chief people officer.

In response to pressure from rivals like DeepSeek, OpenAI is changing the way its o3-mini model communicates its step-by-step “thought” process. ChatGPT users will see an updated “chain of thought” that shows more of the model’s “reasoning” steps and how it arrived at answers to questions. According to a new report from VC firm Andreessen Horowitz (a16z), OpenAI’s AI chatbot, ChatGPT, experienced solid growth in the second half of 2024. ChatGPT’s weekly active users increased to 300 million by December 2024 and 400 million by February 2025. ChatGPT has experienced significant growth recently due to the launch of new models and features, such as GPT-4o, with multimodal capabilities. ChatGPT usage spiked from April to May 2024, shortly after that model’s launch.

It’s complicated and you need to read the small print and anticipate a future state in which a new owner or company takes over responsibility for software licenses. Another recent report concerned Siemens getting into hot water after Broadcom acquired VMware and changed licensing terms. We published a well-read piece pointing out lessons that enterprises can learn from Siemens’ VMware licensing dispute.