25 of the Latest ChatGPT Statistics and Facts in 2023

Microsoft has officially announced the extension of its partnership with OpenAI, the creator of ChatGPT. They will also maintain their exclusive cloud provider status for OpenAI, supporting the training of AI models on Azure.

While the specific financial details were not disclosed, it is rumored that Microsoft's investment could be as much as $10 billion, over several years, improving OpenAI development and ChatGPT usage.

This investment revolved around OpenAI working closely with Azure, Microsoft’s cloud service, making Azure the exclusive cloud service used by OpenAI.

There are many uses to OpenAI’s chatbot. Marketers use it to help them write interesting copy, some people use it to help write a resume, and infamously, students are now using it to write essays.

However, the system has multiple other applications in other industries, leading to other practical ChatGPT trends. For example, 24 healthcare companies are already collaborating with OpenAI, using it to synthesize large volumes of clinical data.

OpenAI and ChatGPT revenue is expected to exceed $1 billion over the next year, primarily from the sale of AI software and the computing capacity that supports it.

This projection is a substantial increase from their initial estimate of $200 million in revenue for this year. The company, supported by Microsoft, is currently generating over $80 million in revenue each month, a significant jump from the $28 million generated in the entire previous year.

OopenAI revenue is up, and the company has experienced a significant increase in its valuation. Thrive Capital is leading a transaction to purchase employee stock, setting OpenAI's value at a minimum of $80 billion.

This development, which was first reported in September, positions OpenAI as one of the world's most valuable private companies. Relatively recently, in Q1 of 2023, OpenAI was valued at approximately $29 billion when it raised $300 million

Sam Altman also admitted that he is “a little bit scared” of the chatbot's potential. He did accept that the alleged artificial intelligence can eliminate certain professions, but also that it can be used for “large-scale disinformation”.

He further claimed that while the chatbot is within human control, he is afraid that some people will ignore the safety limits ChatGPT has, and use it erroneously, or maliciously.

There is already evidence of the bot being significantly biased, since it directly ignores or avoids political questions regarding China and its government. However, when prompted, the bot claimed it does “not self-censor”.

When ChatGPT was asked identical questions, it gave relatively informative answers.

And while there is no official data out yet, GPT-4 costs significantly more.

The cost of using ChatGPT usage is substantial due to the significant computing power required to calculate responses based on user prompts. Namely, the majority of this cost is attributed to the expensive servers needed to support the AI.

Furthermore, the operational costs for using ChatGPT, known as inference costs, have been reported to surpass the training costs when deploying the model at a reasonable scale.

What is surprising is that while ChatGPT is pretty much synonymous with AI chatbots, it is surpassed in downloads by Ask AI and Character AI.

Ask AI, for example, started more than two months before ChatGPT and gained significant traction, especially on Android devices, where it represented over half of all downloads.

On the other hand, Character AI is another AI chatbot application that offers users the ability to customize and personalize AI companions, and reached 8.8 million downloads within 12 days of its release, with the majority of downloads coming from Google Play.

According to a study, participants were aware of the utilization of generative artificial intelligence (AI) in social media, less than a third of them could recognize that they had actually encountered it.

In the United States, the figure was even lower, with fewer than ten percent of respondents having directly engaged with generative AI for social media purposes

The core reason was due to its low accuracy rate. The tool, launched in March 2023, aimed to distinguish between human and machine-written text by assigning probability ratings based on linguistic features.

However, it faced criticism for inaccuracies, particularly in identifying non-native English speakers and mislabeling human-written text as AI-generated.

The famous Microsoft co-founder declared that OpenAI's GPT AI model as the most revolutionary technological advance since the introduction of the graphical user interface in 1980. Drawing a parallel to his own history of shaping Windows around the GUI, Gates now sees GPT models as game-changers.

OpenAI is set to give its ChatGPT chatbot a major update, allowing it to engage in voice conversations, work with images, and expand its capabilities, bringing it closer to AI assistants like Siri.

Additionally, ChatGPT will gain image support, enabling users to seek information about objects in their surroundings.The new features will initially be available to subscribers of the Plus and Enterprise plans by the end of 2023.

OpenAI has reinstated access to ChatGPT in Italy, following privacy-related changes implemented in response to demands from the Italian Data Protection Authority.

These changes include enhancing transparency and making it clearer for European users, including Italian users, to understand how to delete their personal data from the chatbot. The Italian regulator had raised concerns about potential violations of Europe's GDPR privacy law due to data collection practices, prompting OpenAI to update its website with new pages and forms to address these concerns.

A recent analysis reveals that at least 69 out of the top 1,000 most popular websites globally have blocked GPTBot, a web crawler introduced by OpenAI. This number is increasing at a rate of about 5% per week.

Many of these websites, including Amazon, Quora, The New York Times, and others, have opted to block GPTBot, possibly due to concerns about data scraping and a lack of citation or linking to sources by ChatGPT, as well as chatgpt users becoming misinformed.

Even Ivy League schools are not immune to ChatGPT usage.

Namely, in an informal poll 17% of Stanford students admitted to using it, primarily for brainstorming and outlining.

Another poll indicated that a majority of students believe that using ChatGPT for assignments should be considered a violation of the Honor Code, but there are differing opinions on what should be considered a violation.

One of the more worrying ChatGPT and OpenAI trends is the usage in illegal activities.

According to Check Point Research, there is evidence that hackers have been using ChatGPT to improve their malware, as well as to generate code for an infostealer.

Additionally, a hacker named USDoD utilized ChatGPT to create a multi-layer encryption tool, which, although initially benign, could be modified into ransomware to encrypt a user's machine without their consent.

Cybercriminals have also been discussing and demonstrating how ChatGPT can be used to create Dark Web marketplace scripts, facilitating fraudulent activities like selling stolen goods, illegal products, and creating content for illegal distribution.

Chatgpt usage has dropped slightly. In fact, it has experienced a three-month decline in monthly website visits, with a 3.2% drop to 1.43 billion visits in August. The decline follows roughly 10% drops in the preceding two months. Additionally, the time spent by visitors on the website has gradually decreased since March, falling from an average of 8.7 minutes to 7 minutes in August.

However, there are signs of a potential turnaround, as unique visitors worldwide in August slightly increased to 180.5 million users from 180 million. The return of schools in September, with some schools embracing ChatGPT, may boost its traffic and usage, particularly among students seeking homework help.

OpenAI’s chatbot achieved record-breaking numbers, becoming the “fastest-growing consumer application” ever.

For comparison, TikTok took roughly nine months to reach 100 million users, while Instagram had to wait two and a half years.

Furthermore, the UBS study showed that it had, on average, there were 13 million unique ChatGPT users per day (not active users, only visitors) in January of 2022.

Of course, this is to be expected, but an example is in order.

When GPT-3 took an AP Calculus BC exam, it got 20% of answers correct. In comparison, GPT-4 got 80% of them right, and could be classified within the top 10% of test takers.

OpenAI statistics also show that Chat GPT-4 can process up to 25,000 words. In comparison, GPT-3 could handle roughly 3,000 to 4,000.

ChatGPT stats show just how insanely popular this program is. However, sometimes the content it provides, while sounding authoritative, is actually filled with incorrect information. Furthermore, AI image-creation tools can also lead to the promotion of misinformation, and even plagiarism.

For these reasons, TikTok has implemented guidelines that require some sort of “disclaimer”, through a sticker or a caption that warns users that the content they are seeing is AI-generated.

DALL-E is a product that is just as fascinating as ChatGPT but only gets a fraction of the attention. It's a generative language model that lets you create images based on text input. OpenAI stats show that DALL-E 2 has received a substantial upgrade, and can provide images with four times better resolution, and in less time, when compared to its predecessor.

OpenAI stats show many healthcare companies already partnering up with OpenAI. And some companies have recognized the potential of OpenAI’s DALL-E 2 product in radiology.

In fact, according to a paper recently published in the Journal of Medical Internet Research, DALL-E 2 was capable of creating realistic x-ray images through simple text prompts. In time, with (much more) development, it might be able to create a “complete full-body radiograph” by only being fed an image of a patient’s knee.

Until very recently, one of the more annoying ChatGPT facts was that it had no real time internet access. In September of 2023, that changed.

OpenAI introduced a real-time internet access feature, allowing ChatGPT to provide direct links to online sources. Initially available in beta for ChatGPT Plus and Enterprise subscribers, this feature enables access to up-to-date information beyond 2021, enhancing the chatbot's capabilities and reliability.

ChatGPT stats and data show the chatbot has been trained on a significant amount of data. However, the quality, and content age rating, of said data is questionable.

BlueMail, which is hosted on Apple’s app store, recently announced an update that integrates the app with ChatGPT. However, this update does not have content filtering.

For this reason, Apple has blocked the app's update, until it puts an age restriction with 17 and older.