The Rise and Evolution of ChatGPT

ChatGPT, OpenAI's revolutionary text-generating AI chatbot, has taken the world by storm since its launch in November 2022. Initially a productivity tool for writing essays and code with short prompts, it has evolved into a behemoth used by over 92% of Fortune 500 companies. This growth has propelled OpenAI itself into one of the most-hyped companies in recent memory.

OpenAI's latest partnership with Apple for its upcoming generative AI offering, Apple Intelligence, has further solidified OpenAI's position in the AI race.

A Timeline of Recent ChatGPT Updates

The year 2024 has seen a flurry of updates and releases for ChatGPT. This section dives into the key highlights, starting with the most recent.

September 2024

September witnessed several key milestones for ChatGPT.

• OpenAI rolled out Advanced Voice Mode, expanding its availability to a wider range of paying customers. This mode features a new design, improved conversational speed, enhanced accents in foreign languages, and five new voices.
• A YouTuber showcased a hack enabling ChatGPT to run on a TI-84 graphing calculator, sparking concerns about potential cheating in schools.
• OpenAI unveiled a preview of OpenAI o1, also known as "Strawberry," a new model capable of self-fact-checking and enhanced reasoning in math and science. OpenAI plans to make o1-mini accessible to all free users in the future.
• A hacker successfully tricked ChatGPT into generating instructions for making explosives, highlighting the potential risks of AI's misuse.
• OpenAI announced exceeding 1 million paid users across its corporate offerings, including ChatGPT Team, ChatGPT Enterprise, and ChatGPT Edu.
• Volkswagen is rolling out its ChatGPT voice assistant to the US market, integrating the AI chatbot into its vehicles.

August 2024

OpenAI continued its aggressive pace of development in August.

• OpenAI inked a content deal with Condé Nast, allowing the company to surface stories from publications like The New Yorker and Vogue within ChatGPT and SearchGPT.
• TechCrunch's Maxwell Zeff provided early insights into OpenAI's Advanced Voice Mode, praising its speed and ability to answer complex questions.
• OpenAI shut down a ChatGPT-powered Iranian influence operation aiming to generate content about the U.S. presidential election.
• OpenAI discovered that GPT-4o, the model powering Advanced Voice Mode, can exhibit strange behaviors, such as mimicking voices or shouting randomly.
• ChatGPT's mobile app experienced its highest revenue month, reaching $28 million, driven by the release of the GPT-4o model.
• OpenAI is developing a watermarking tool to detect cheating by students using ChatGPT, but its release remains under consideration.

July 2024

OpenAI made significant advancements in July, pushing the boundaries of AI capabilities.

• OpenAI released the alpha version of GPT-4o's Advanced Voice Mode to a select group of ChatGPT Plus users. This feature, initially delayed due to controversy regarding the voice's resemblance to Scarlett Johansson, is expected to become widely available in the fall.
• OpenAI announced SearchGPT, a new AI search experience designed to compete with Google, offering timely answers and follow-up question capabilities. This prototype is currently undergoing testing with a limited group of users.
• Reports suggest that OpenAI could lose up to $5 billion in 2024 due to the high cost of operating ChatGPT.
• OpenAI unveiled GPT-4o mini, a smaller and faster AI model aimed at replacing GPT-3.5 Turbo as the smallest model offered.
• OpenAI partnered with the Los Alamos National Laboratory to explore the use of AI in healthcare and bioscience research.

June 2024

June saw OpenAI expanding its partnerships and addressing concerns regarding ChatGPT's capabilities.

• OpenAI announced CriticGPT, a model trained on GPT-4, designed to find errors in ChatGPT's code output and assist in improving the model's performance.
• OpenAI inked a content deal with TIME, providing the magazine's content to ChatGPT and gaining access to OpenAI's technology for developing new products.
• OpenAI delayed the launch of Advanced Voice Mode to July due to ongoing technical issues, with a full rollout to ChatGPT Plus users potentially taking place in the fall.
• ChatGPT released an app for macOS, providing a desktop-based interface for interacting with the chatbot and offering additional features.
• Apple announced at WWDC 2024 that it would be integrating ChatGPT into Siri and other first-party apps across its operating systems. These integrations, powered by GPT-4o, will be available on iOS 18, iPadOS 18, and macOS Sequoia later this year.
• The House Oversight subcommittee invited Scarlett Johansson to testify about the controversy surrounding OpenAI's Sky voice, raising concerns about deepfakes.
• ChatGPT experienced two outages in a single day, highlighting the challenges of maintaining reliable service for a rapidly growing platform.

May 2024

May saw OpenAI continue to expand its reach and address controversies while releasing its latest flagship model.

• The Atlantic and Vox Media announced licensing and product partnerships with OpenAI, providing access to their content for use in ChatGPT.
• OpenAI signed a deal with PwC to provide ChatGPT Enterprise access to 100,000 employees, becoming OpenAI's largest customer to date.
• OpenAI confirmed that it has begun training its successor to GPT-4, hinting at the continued evolution of its AI capabilities.
• Former OpenAI board member Helen Toner revealed that the board was not aware of ChatGPT's launch until it happened, raising questions about OpenAI's communication and governance practices.
• ChatGPT's mobile app revenue saw a significant surge following the release of GPT-4o, demonstrating the model's popularity among mobile users.
• OpenAI announced the pausing of its Sky voice, which had been deemed too similar to Scarlett Johansson's voice in "Her," highlighting the ethical considerations in AI voice technology.
• OpenAI added new features for data analysis within ChatGPT, including file uploads from Google Drive and Microsoft OneDrive, table and chart interaction, and chart export capabilities.
• OpenAI announced a partnership with Reddit, providing the company with access to real-time data from the social network for training ChatGPT and developing new features.
• OpenAI debuted GPT-4o, its new omni model, which features a black hole-like interface and capabilities for speech, image, and vision. GPT-4o is being rolled out gradually across OpenAI's products.
• OpenAI is building a tool called Media Manager to allow content creators to control how their content is used for training AI models and opt out if they wish.
• OpenAI is exploring the possibility of allowing explicit images and text in its AI products, raising ethical questions about its handling of sensitive content.
• OpenAI partnered with Stack Overflow, providing access to its API and receiving feedback from developers to improve its AI models.

April 2024

April saw OpenAI face legal challenges while expanding into new markets and refining its model.

• Alden Global Capital-owned newspapers filed a copyright lawsuit against OpenAI and Microsoft, alleging the companies used copyrighted articles without permission.
• OpenAI entered into a content licensing deal with the Financial Times, marking another step in its effort to leverage real-world data for training its AI models.
• OpenAI opened a new office in Tokyo and announced plans for a GPT-4 model optimized for the Japanese language.
• Sam Altman pitched ChatGPT Enterprise to Fortune 500 companies, highlighting its potential for corporate use.
• OpenAI released an updated version of GPT-4 Turbo for premium users, claiming improvements in writing, math, reasoning, and coding.
• OpenAI made ChatGPT accessible without requiring an account, but with limitations on functionality and content policies.

March 2024

March saw OpenAI navigate controversies related to data privacy, AI's environmental impact, and the ethical considerations of using AI for content generation.

• OpenAI's GPT Store faced concerns about spam and potentially copyright-infringing GPTs.
• The New York Times disputed OpenAI's claims about "hacking" its products in response to the copyright lawsuit.
• OpenAI's VP, Peter Deng, declined to comment on whether artists should be compensated for their work being used to train generative AI models.
• A report estimated that ChatGPT uses an enormous amount of electricity daily, highlighting its environmental impact.
• OpenAI added a Read Aloud feature to the web version of ChatGPT, allowing users to listen to its responses in multiple languages.

February 2024

February saw OpenAI expanding its reach into new markets and facing scrutiny for its use of AI in various applications.

• OpenAI partnered with Dublin City Council to use GPT-4 for tourism purposes, developing personalized itineraries for travelers.
• A law firm was criticized by a judge for using ChatGPT to justify its hourly billing rate, raising concerns about the use of AI in legal services.
• ChatGPT experienced a bug that resulted in users receiving nonsensical answers for several hours, demonstrating the potential for technical issues on a large scale.
• Match Group announced a deal with OpenAI, using ChatGPT to help its employees with work-related tasks.
• OpenAI began rolling out new "memory" controls for ChatGPT, allowing users to control what the chatbot remembers and forgets.
• OpenAI introduced the "Temporary Chat" feature, enabling users to have conversations with a clean slate, but with potential for data retention.

January 2024

January saw OpenAI facing regulatory scrutiny and ethical dilemmas, while also introducing new features and partnerships.

• Paid ChatGPT users gained the ability to directly invoke GPTs within conversations, allowing for greater flexibility in utilizing different AI models.
• Reports emerged that ChatGPT might be leaking usernames and passwords from users' private conversations, raising concerns about data security.
• Italy's data protection authority investigated ChatGPT, concluding that it potentially violated EU privacy laws, leading to a temporary ban on the service in the country.
• OpenAI partnered with Common Sense Media to develop AI guidelines and education materials for parents and educators.
• OpenAI responded to concerns from the Congressional Black Caucus regarding the lack of diversity on its board, committing to building a more inclusive leadership team.
• OpenAI lowered prices for its API and addressed issues with GPT-4's "laziness," aiming to improve the model's responsiveness.
• OpenAI banned a developer who created a chatbot impersonating a presidential candidate, highlighting its commitment to combating election misinformation.
• OpenAI announced a partnership with Arizona State University, making ChatGPT Enterprise available to the institution for educational and research purposes.
• The winner of a prestigious literary prize revealed that she used ChatGPT to generate parts of her novel, sparking debates about AI's role in creative writing.
• Sam Altman hinted at the upcoming release of GPT-5, which will feature multimodal capabilities for speech, image, code, and video.
• OpenAI announced the formation of a Collective Alignment team to collect public input on its models' behavior, promoting ethical AI development.
• OpenAI unveiled a plan to combat election misinformation, including banning the use of its tools for political campaigning and lobbying.
• OpenAI made a policy change to allow military applications, provided they adhere to its guidelines on responsible AI use.
• OpenAI launched ChatGPT Team, a subscription plan aimed at small teams, providing a dedicated workspace and administrative tools.
• OpenAI officially launched the GPT Store, a marketplace for custom AI-powered chatbots accessible to premium ChatGPT users.
• OpenAI argued that training AI models without copyrighted materials would be "impossible" in response to proposed regulations in the UK.
• OpenAI publicly responded to The New York Times' copyright lawsuit, claiming that the case is without merit.
• OpenAI's GPT Store was scheduled to launch in early January, offering a platform for developers to share their GPTs.
• OpenAI made changes to its terms of service, including establishing OpenAI Ireland Limited as the data controller for users in the EEA and Switzerland, aimed at mitigating regulatory risks in the EU.

FAQs

This section provides answers to common questions surrounding ChatGPT.

What is ChatGPT? How does it work?

ChatGPT is an AI-powered chatbot developed by OpenAI. It utilizes a large language model called GPT-4, which is trained on vast amounts of text data, enabling it to generate human-like text in response to prompts. It uses deep learning to analyze patterns and create coherent and relevant responses.

When was ChatGPT released?

ChatGPT was released for public use on November 30, 2022.

What is the latest version of ChatGPT?

Both the free and paid versions of ChatGPT are regularly updated. The most recent model is GPT-4o, released in May 2024.

Can I use ChatGPT for free?

Yes, there is a free version of ChatGPT available, requiring only a sign-in. There is also a paid version, ChatGPT Plus, which offers additional features.

Who uses ChatGPT?

Anyone can use ChatGPT! It is widely adopted by individuals, businesses, and organizations for various purposes, including writing, coding, research, and customer service.

What companies use ChatGPT?

Many enterprises utilize ChatGPT, including Microsoft, which has integrated it into its Bing search engine and Windows 11. Other companies, like Looking Glass and Solana, leverage ChatGPT for unique applications.

What does GPT mean in ChatGPT?

GPT stands for Generative Pre-trained Transformer.

What is the difference between ChatGPT and a chatbot?

A chatbot is any software that interacts with users through dialogue. ChatGPT is a specific type of chatbot that utilizes AI and a large language model for more advanced and nuanced responses.

Can ChatGPT write essays?

Yes, ChatGPT can generate essays. However, it's crucial to use it responsibly and avoid plagiarism.

Can ChatGPT commit libel?

ChatGPT can generate false or harmful statements, which could potentially lead to libel claims. OpenAI is working on mitigating these risks but emphasizes that users must act responsibly when using the technology.

Does ChatGPT have an app?

Yes, there is a free ChatGPT mobile app available for iOS and Android devices.

What is the ChatGPT character limit?

While there is no officially documented character limit, users have observed limitations after around 500 words.

Does ChatGPT have an API?

Yes, OpenAI released an API for ChatGPT on March 1, 2023, allowing developers to integrate the chatbot into their applications.

What are some everyday uses for ChatGPT?

Everyday uses include writing emails, generating blog ideas, summarizing text, creating shopping lists, and composing creative content.

What are some advanced uses for ChatGPT?

Advanced uses include debugging code, generating complex programs, exploring scientific concepts, solving complex problems, and conducting research.

How good is ChatGPT at writing code?

ChatGPT can generate code in various languages, but it may lack context awareness and might not always produce code that is fully functional or appropriate for specific situations.

Can you save a ChatGPT chat?

Yes, ChatGPT allows users to save their chats in the interface, accessible in the sidebar.

Are there alternatives to ChatGPT?

Yes, several other AI-powered chatbots exist, including Google's Gemini, Anthropic's Claude, and various open-source alternatives.

How does ChatGPT handle data privacy?

OpenAI offers a form for individuals to object to the processing of their personal information by its AI models. It also acknowledges the importance of data privacy while balancing it with freedom of expression.

What controversies have surrounded ChatGPT?

ChatGPT has been involved in several controversies, including the generation of potentially illegal instructions by Discord's chatbot, false claims that led to potential defamation lawsuits, and concerns about plagiarism in educational settings.

Where can I find examples of ChatGPT prompts?

Various marketplaces offer ChatGPT prompts, such as PromptBase and ChatX.

Can ChatGPT be detected?

Current tools for detecting ChatGPT-generated text are not entirely reliable, with varying degrees of accuracy.

Are ChatGPT chats public?

No, ChatGPT chats are not public. However, a bug that exposed conversation titles to other users was recently fixed.

What lawsuits are there surrounding ChatGPT?

While no lawsuits target ChatGPT specifically, the legal battles surrounding OpenAI's use of data for training AI models have broader implications for ChatGPT.

Are there issues regarding plagiarism with ChatGPT?

Yes, plagiarism is a concern because ChatGPT can regurgitate content from its training data, potentially leading to accidental or intentional copying.