OpenAI’s text-generating AI chatbot, ChatGPT, has taken the globe by storm. More than 92% of Fortune 500 companies now use what started out as a tool to increase productivity by having users write essays and code in response to brief text prompts. And that growth has catapulted OpenAI to become one of the most heralded startups in recent memory, despite the fact that CEO and co-founder Sam Altman’s ouster and quick return raised questions about the company’s direction and opened the way for rivals.
What does this entail for OpenAI, ChatGPT, and other projects? The fallout is still settling, but it might help rivals like Meta and its LLaMA family of big language models get attention and investment while the industry watches OpenAI crumble and rebuild.
While ChatGPT has a more…nefarious side, it’s evident that AI tools aren’t going away anytime soon. ChatGPT has reached 100 million weekly active users since its original introduction over a year ago, and OpenAI is significantly investing in it.
Prior to the leadership crisis, OpenAI had its inaugural developer conference, OpenAI DevDay, on November 6. It revealed a plethora of GPT improvements at the conference, including GPT-4 Turbo (supercharged versions of GPT-4, its newest language-writing model), a multimodal API, and a GPT store where users can develop and commercialize their own bespoke versions of GPT.
GPT-4, which can write more organically and eloquently than earlier models, is generally available only to ChatGPT subscribers who pay. GPT-4, on the other hand, is available for free via Microsoft’s Bing Chat in Microsoft Edge, Google Chrome, and Safari web browsers. Aside from the launches of GPT-4 and OpenAI DevDay, OpenAI recently opened ChatGPT to the internet for all users. Users may now produce text prompts and graphics directly in ChatGPT thanks to the integration of DALL-E 3.
Here’s a history of ChatGPT product changes and releases, beginning with the most recent and progressing through the year. If you have any more questions, please see our ChatGPT FAQ.
December 20, 2023
OpenAI has reopened subscriptions to ChatGPT Plus.
After suspending ChatGPT Plus subscriptions in November owing to a “surge in usage,” OpenAI CEO Sam Altman stated that sign-ups are now open again. GPT-4 and GPT-4 Turbo are included with the Plus membership.
OpenAI and Axel Springer collaborate on “real-time” ChatGPT news.
OpenAI has entered into a new agreement with Berlin-based news publisher Axel Springer, which owns Business Insider and Politico, to “help provide people with new ways to access quality, real-time news content through our AI tools.” Under the terms of the agreement, OpenAI will train its generative AI models on the publisher’s content and add recent Axel Springer-published articles to ChatGPT.
According to Stanford academics, ChatGPT did not result in an increase in high school cheating.
According to a new Stanford University study, the use of chatbots such as ChatGPT has not resulted in an increase in cheating across US high schools. Cheating rates are comparable across the board this year, according to a poll of more than 40 US high schools.
Users of ChatGPT are concerned that the chatbot is suffering from seasonal depression.
Users of ChatGPT have reported that the chatbot has been “lazier” than usual since November, noting instances of easier responses and refusal to execute specified tasks. OpenAI has acknowledged that they are aware of the problem but are unsure why it is occurring.
Some people believe it supports the “winter break hypothesis,” which contends that AI is poorer in December because it “learned” to perform less work over the holidays, while others speculate that the chatbot is faking seasonal sadness.
ChatGPT is now permitted in judicial verdicts by UK judges.
The UK Judicial Office published instructions allowing judges to draft legal opinions and conduct court functions using ChatGPT and other AI technologies. The guideline outlines how to utilize AI properly in the courts, such as being mindful of possible prejudice and protecting privacy.
Following the Google DeepMind test, OpenAI has made repeating sentences “forever” a violation of its terms of service.
Following a Google DeepMind experiment in which ChatGPT repeated chunks of its training data, OpenAI has reported requesting ChatGPT to repeat particular phrases “forever” as a breach of its terms of service.
Brazilian legislators pass an ordinance drafted by ChatGPT.
City legislators in Brazil unknowingly passed laws totally authored by ChatGPT. Porto Alegre councilman Ramiro Rosário revealed weeks after the measure was enacted that he utilized ChatGPT to create the proposal and did not inform fellow council members until after the fact.
According to reports, OpenAI has postponed the opening of its GPT store until 2024.
According to an internal letter obtained by Axios, OpenAI intends to postpone the debut of its much-anticipated GPT store until early 2024. Custom GPTs and the associated shop were a key announcement at OpenAI’s DevDay conference, with the store set to debut last month.
November 20, 2023
ChatGPT’s mobile applications have over 110 million installations and approximately $30 million in revenue.
According to data.ai, after releasing for iOS and Android in May and July, ChatGPTs have surpassed 110 million combined installations and approximately $30 million in user spending.
ChatGPT celebrates its one-year milestone.
OpenAI has reached a significant milestone: one year of ChatGPT. What started as a “low-key research preview” grew into a behemoth that profoundly altered the AI business. In a post on X, CEO Sam Altman reflected on the night before the launch, saying, “What a year it’s been…”
ChatGPT is not named as an ‘app of the year’ by Apple or Google.
Despite ChatGPT’s mobile app becoming the fastest-growing consumer application in history until Meta’s Threads overtook it, neither Apple nor Google chose an AI app as their app of the year for 2023.
ChatGPT is forced to provide training data due to a researcher’s assault.
Researchers from Google DeepMind found a sizable amount of personally identifying information in OpenAI’s LLMs. The test comprised asking ChatGPT to repeat the word “poem” indefinitely, which eventually led to the chatbot churning out sensitive information like email addresses and phone numbers.
ChatGPT and other AI chatbots are increasing the number of phishing emails.
According to recent SlashNext research, there has been a 1,265% rise in malicious phishing emails since the fourth quarter of 2022. According to the survey, fraudsters are heavily using AI technologies, such as ChatGPT, to create enticing and sophisticated phishing emails.
South African authorities are looking into whether President Cyril Ramaphosa used ChatGPT to compose a speech.
Following conjecture, social media users ran chunks of Ramaphosa’s November 21 speech in Johannesburg through AI detectors, claiming that some of it was authored using ChatGPT. Local authorities are looking into the accusations after South African presidential spokesman Vincent Magwenya rejected them.
Siri may be replaced with ChatGPT Voice.
With the release of OpenAI’s ChatGPT Voice functionality to all free users, it can now be used to replace Siri on an iPhone 15 Pro and Pro Max by setting the new Action Button. The new feature allows you to ask ChatGPT questions and listen to its replies, much like a much better Siri.
Sam Altman is back as CEO.
Altman’s return was rapid, with an “agreement in principle” revealed between him and OpenAI’s board that would see him reinstated as CEO and the board restructured to include new members, including former U.S. Treasury Secretary Larry Summers. The most important lesson for ChatGPT is that the members of the board who are more concerned with the commercialization of OpenAI’s products have been pushed to the side.
ChatGPT Voice is now available to all free users.
Even though its leadership is changing, OpenAI continues to provide improvements to ChatGPT. The text-to-speech approach, first unveiled in September and made available to paying customers on a rolling basis, can generate a voice from text instructions and a few seconds of speech samples. By accessing the settings on your mobile ChatGPT apps and clicking the “headphones” icon, you can test out the five voice options that OpenAI created in collaboration with voice actors.
Sam Altman may return, but the situation is difficult.
The only constant at OpenAI right now is change, and in a series of interviews, Nadella tempered previous reports that Altman and Brockman were on their way to Microsoft.
“Obviously, we want Sam and Greg to have a fantastic home if they’re not going to be in OpenAI,” Nadella told CNBC, adding that we were “open” to them staying at Microsoft or returning to OpenAI if the board and staff supported the move.
Confirmation that Sam Altman would not be returning as CEO of OpenAI.
After his surprise sacking by the company’s board, a number of investors and OpenAI staff attempted to bring Altman back, but after a weekend of discussions, it was revealed that Altman would not return to OpenAI and that new leadership would take over. What this implies for the future of ChatGPT and the OpenAI Dev Day announcements remains to be seen.
Sam Altman has been fired as CEO of OpenAI.
OpenAI’s Sam Altman has been dismissed. He will step down as CEO and quit the company’s board of directors, with OpenAI’s chief technical officer, Mira Murati, taking over as temporary CEO. According to an OpenAI blog post, the board “no longer has confidence in [Altman’s] ability to continue leading OpenAI.”
Altman stated in a statement on X that working with OpenAI “was transformative” for him and “hopefully the world.”
OpenAI investigates the application of ChatGPT in the classroom.
At a conference in San Francisco, OpenAI COO Brad Lightcap disclosed that the business would likely form a team to investigate how AI and ChatGPT might be utilized in education. This statement comes at a time when educators are criticizing ChatGPT for inciting cheating, leading to bans in several school districts.
Due to a “surge in usage,” OpenAI has paused new ChatGPT Plus subscriptions.
Following OpenAI’s Dev Day presentation, Sam Altman said that new subscriptions for the company’s premium ChatGPT Plus product would be paused. The temporary suspension of sign-ups, as well as the desire for ChatGPT Plus’ additional capabilities, such as creating bespoke GPTS, has resulted in a wave of eBay resellers.
ChatGPT is marked as possibly hazardous to children.
According to an independent assessment by the nonprofit advocacy group Common Sense Media, ChatGPT may be risky for younger users. The research gave ChatGPT a three-star rating overall, with the lowest grades pertaining to openness, privacy, trust, and safety.
OpenAI blames the ChatGPT outage on a DDoS assault.
OpenAI claimed that a DDoS attack was to blame for the disruptions affecting ChatGPT and its developer tools. For around 24 hours, ChatGPT had occasional outages, preventing users from logging in or using the service.
OpenAI introduces the GPT-4 Turbo.
During its first OpenAI DevDay conference, OpenAI revealed GPT-4 Turbo. GPT-4 Turbo is available in two versions: one that just analyzes text and another that recognizes the context of both text and graphics.
GPT-4 is fine-tuned.
Unlike the fine-tuning approach for GPT-3.5, the GPT-4 program will include additional monitoring and direction from OpenAI teams, according to the firm, owing to technological challenges.
OpenAI’s GPT Store allows you to create (and market) your own GPT.
Users and developers will soon be able to create their own GPT without any coding knowledge. Anyone who creates their own GPT will be able to list and monetize it on OpenAI’s marketplace in the future.
ChatGPT has 100 million active weekly users.
ChatGPT boasts 100 million weekly active users after being published about a year ago. Over two million developers, including more than 92% of Fortune 500 companies, are using the platform, according to OpenAI CEO Sam Altman.
OpenAI introduces the DALL-E 3 API as well as new text-to-speech models.
After debuting on ChatGPT-4 and Bing Chat, OpenAI’s text-to-image model, DALL-E 3, is now accessible via API. The Audio API, OpenAI’s recently released text-to-speech API, has six preset voices and two generative AI model versions.
OpenAI offers to defend commercial clients against copyright lawsuits.
In response to peer pressure, OpenAI has agreed to reimburse the legal fees of clients who have been sued over intellectual property claims based on work done by an OpenAI service. The safeguards do not seem to apply to all OpenAI products, including the free and Plus levels of ChatGPT.
As OpenAI’s multimodal API becomes more widely available, research demonstrates that it is still faulty.
GPT-4 with Vision will be made public by OpenAI with the future release of the GPT-4 Turbo API, although several researchers discovered that the model is still defective in numerous important and serious ways.
OpenAI has released an API that allows developers to include ‘assistants’ in their products.
OpenAI launched the Assistants API at its OpenAI DevDay to assist developers in creating “agent-like experiences” inside their applications. Natural language-based data analysis software, coding assistance, or even an AI-powered holiday planner are examples of use cases.
October 20, 2023
ChatGPT app income is not declining, although it is not the highest.
In terms of downloads, OpenAI’s chatbot app vastly outpaces all others on mobile devices, but it is unexpectedly not the top AI app in terms of income. Several other AI chatbots, such as “Chat & Ask AI” and “ChatOn—AI Chat Bot Assistant,” are actually producing more money than ChatGPT.
ChatGPT evaluates Plus users’ capacity to upload and analyze files.
ChatGPT Enterprise Plan subscribers have reported additional beta capabilities, such as the ability to submit PDFs for analysis and immediately ask questions about them. When utilizing ChatGPT, customers no longer have to explicitly pick a mode, such as DALL-E or surfing, thanks to the new rollout. Instead, depending on the question, consumers will be automatically converted to models.
ChatGPT now has a web search.
OpenAI has publicly reintroduced its internet-browsing capability in ChatGPT, three weeks after reintroducing it in beta following a gap of many months. The AI chatbot had previously been restricted to data until September 2021.
OpenAI has now integrated DALL-E 3 into ChatGPT.
Because of the connection, users no longer have to think about their text prompts when requesting DALL-E to make a picture. Users will also be able to obtain graphics in response to text-based requests without switching between applications.
GPT-4 issues discovered by Microsoft-affiliated research
A Microsoft-affiliated research article investigated the “trustworthiness”—and toxicity—of LLMs such as GPT-4. Because GPT-4 is more likely to obey the directions of “jailbreaking” prompts, the co-authors suggest that GPT-4 may be led to regurgitate poisonous, prejudiced language more readily than other LLMs.
ChatGPT’s mobile app generated a record $4.58 million in revenue in September.
In September, OpenAI had 15.6 million downloads and approximately $4.6 million in gross revenue across its iOS and Android applications globally. According to recent statistics from market intelligence company Appfigures, revenue growth has already started to decline, dipping from 30% to 20% in September.
September 20, 2023
ChatGPT can now access the internet again.
OpenAI announced on Twitter/X that ChatGPT may now access the internet and is no longer restricted to data collected before September 2021. Back in July, the chatbot could browse the web for Plus customers, but the functionality was removed when users abused it to circumvent paywalls.
ChatGPT has a voice now.
OpenAI has announced the addition of a new voice for verbal interactions as well as image-based smarts to its AI-powered chatbot.
Poland launches a probe into OpenAI.
The Polish authorities declared publicly that they had begun an inquiry against ChatGPT, accusing the business of a series of violations of the EU’s General Data Protection Regulation (GDPR).
OpenAI unveils DALL-E 3.
ChatGPT is used to assist in filling in prompts in the improved text-to-image tool DALL-E 3. Subscribers to OpenAI’s premium ChatGPT subscriptions, ChatGPT Plus and ChatGPT Enterprise, may enter in an image request and refine it via discussions with the chatbot, obtaining results right inside the chat app.
ChatGPT-powered AI is integrated into Opera GX.
Aria, an AI browser powered by OpenAI’s ChatGPT, debuted on Opera in May to make it simpler for users to search, ask questions, and write code. Today, the business announced the addition of Aria to Opera GX, a version of the main Opera browser designed just for gamers.
Users of Opera GX may now engage directly with a browser AI to discover the most recent gaming news and advice.
August 20, 2023
OpenAI has published a handbook for instructors who use ChatGPT in the classroom.
OpenAI hopes to improve the system’s reputation in the classroom, since ChatGPT has proved contentious in the classroom owing to plagiarism. OpenAI has provided a number of options for putting the chatbot to use in the classroom.
ChatGPT Enterprise is now available via OpenAI.
ChatGPT Enterprise can do the same things as ChatGPT, such as write emails, write essays, and debug computer code. On top of the vanilla ChatGPT, the new service offers “enterprise-grade” privacy and data analysis capabilities, as well as improved speed and customization possibilities.
According to a survey, only a small percentage of Americans utilize ChatGPT.
According to a recent Pew study, the language model isn’t as popular or as dangerous as some would have you believe. According to ongoing Pew Research surveys, although ChatGPT is increasing mindshare, only around 18% of Americans have ever used it.
GPT-3.5 Turbo now has fine-tuning thanks to OpenAI.
Companies employing GPT-3.5 Turbo via the company’s API may fine-tune the model to better obey particular directions. For instance, having the model always reply in a certain language. Or enhancing the model’s capacity to structure replies consistently, as well as honing the “feel” of the model’s output, such as tone, to better suit a brand or voice. Most significantly, fine-tuning allows OpenAI users to reduce the length of text prompts in order to speed up API requests and save money.
OpenAI is collaborating with Scale AI to enable businesses to fine-tune GPT-3.5. However, it is unclear if OpenAI is building an in-house tuning tool to complement platforms like Scale AI or for a whole other reason.
Costs of fine-tuning:
Training costs $0.008 for every 1,000 tokens.
Input: $0.012 per 1000 tokens
Output: $0.016 per 1000 tokens
OpenAI has acquired Global Illumination.
OpenAI announced its first public purchase in its seven-year existence: Global Illumination, a New York-based firm that uses AI to develop creative tools, infrastructure, and digital experiences.
“We’re very excited for the impact they’ll have here at OpenAI,” the company said in a short statement on its official blog. “The entire team has joined OpenAI to work on our core products, including ChatGPT.”
The ‘custom instructions’ functionality is now available to ChatGPT-free users.
OpenAI has announced that bespoke instructions will be available to all users, even those on the free tier of service. Users may utilize this feature to specify whatever preferences and criteria they want the AI chatbot to take into account while answering.
AI applications in China must receive an administrative license.
Several generative AI applications have been withdrawn from Apple’s China App Store in advance of the country’s current generative AI laws, which go into force on August 15.
“As you may be aware, the government has tightened laws governing deep synthesis technologies (DST) and generative AI services such as ChatGPT. “To operate in China, DST must meet permitting requirements, including obtaining a license from the Ministry of Industry and Information Technology (MIIT),” Apple said in a letter to OpenCat, a native ChatGPT client. “Based on our review, your app is associated with ChatGPT, which does not have the requisite permits to operate in China.”
July 2023
ChatGPT for Android has now been released in the United States, India, Bangladesh, and Brazil.
OpenAI has launched ChatGPT for Android only a few days after setting up a purchase page on Google Play. The app is currently available in a few countries.
ChatGPT is making its way to Android.
ChatGPT is now available for “pre-order” for Android users.
The Android ChatGPT app seems to be functionally equivalent to the iOS version, which means it has access to the most, if not all, of the web-based version’s capabilities. You should be able to sync your chats and settings between devices as well, so there are no concerns if you have an iPhone at home and an Android at work.
OpenAI introduces ChatGPT-specific instructions.
OpenAI introduced bespoke instructions for ChatGPT users so that they do not have to send the same instruction prompts to the chatbot every time they engage with it.
This feature, according to the business, allows users to “share anything you’d like ChatGPT to consider in its response.” For example, a teacher may state they teach fourth-grade arithmetic, while a developer can identify the coding language they prefer when asking for advice. A user may also define their family size so that the text-generating AI can provide appropriate replies on meals, grocery shopping, and vacation planning.
According to reports, the FTC is looking into OpenAI.
According to reports, the FTC is investigating whether OpenAI’s flagship ChatGPT conversational AI made “false, misleading, disparaging, or harmful” claims about individuals.
Devin Coldewey of Eltrys reports:
The FTC doesn’t simply glance around and say, “That looks suspicious.” Typically, a lawsuit or official complaint is brought to their attention, and the activities stated in it signal that rules are being violated. For example, a person may sue a supplement business because the pills got them sick, and the FTC will investigate since there is proof the firm misled about the adverse effects.
GPT-4 has been made generally available via OpenAI.
GPT-4 will be available to all current OpenAI developers “with a history of successful payments” beginning July 6. OpenAI intends to provide new developers with access by the end of July.
OpenAI adds that in the future, developers will be able to fine-tune GPT-4 and GPT-3.5 Turbo, two of the original models powering ChatGPT, using their own data, as they have with many of OpenAI’s other text-generating models. According to OpenAI, this functionality will be available later this year.
June 2023
The ChatGPT app can now only search the web on Bing.
OpenAI stated that ChatGPT Plus customers may now access a new tool called Browsing, which enables ChatGPT to search Bing for answers to inquiries.
To activate the Browsing function, go to the app settings’ New Features section, choose “GPT-4” in the model switcher, and then pick “Browse with Bing” from the drop-down list. Both the iOS and Android ChatGPT applications support browsing.
Mercedes is integrating ChatGPT into its infotainment system.
Owners of Mercedes-Benz cars equipped with MBUX in the United States will be able to enroll in a beta program beginning June 16 to activate the ChatGPT capability. The extremely adaptable big language model will be able to supplement the car’s communication abilities as a result. You just tell your automobile, “Hey Mercedes, I want to join the beta program.”
But it’s not quite apparent why.
The ChatGPT app is now available for iPad, with Siri and Shortcuts compatibility.
The latest ChatGPT app version adds native iPad compatibility as well as the ability to use the chatbot with Siri and shortcuts. Users may now move individual messages from ChatGPT into other applications using drag-and-drop.
ChatGPT now operates in full-screen mode on the iPad, which is suited for the tablet’s interface.
May 2023
A Texas court has ordered that any AI-generated material be reported and vetted.
The federal judge in Texas has imposed a requirement that any attorney who appears in his court swear that “no portion of the filing was drafted by generative artificial intelligence,” or that if it was, it was verified “by a human being.”
The ChatGPT app is now available in over 30 countries.
Algeria, Argentina, Azerbaijan, Bolivia, Brazil, Canada, Chile, Costa Rica, Ecuador, Estonia, Ghana, India, Iraq, Israel, Japan, Kazakhstan, Kuwait, Lebanon, Lithuania, Mauritania, Mauritius, Mexico, Morocco, Namibia, Nauru, Oman, Pakistan, Peru, Poland, Qatar, Slovenia, Tunisia, and the United Arab Emirates are among the new countries.
The ChatGPT app is now accessible in 11 more countries.
OpenAI said in a tweet that the ChatGPT mobile app is now accessible on iOS in the United States, Europe, South Korea, and New Zealand, and that more people will be able to download it shortly. The software received over 500,000 downloads in only six days.
The ChatGPT iOS app is now accessible in 11 more countries: Albania, Croatia, France, Germany, Ireland, Jamaica, Korea, New Zealand, Nicaragua, Nigeria, and the United Kingdom. More will be added shortly!
May 24, 2023—OpenAI (@OpenAI)
OpenAI has released a ChatGPT app for iOS.
ChatGPT is now available on mobile devices. According to the firm, the new ChatGPT app will be free to use, ad-free, and will include voice input, although it will initially be restricted to customers in the United States.
When you use the mobile version of ChatGPT, it will sync your history across devices, which means it will remember what you previously looked for via its online interface and make it available to you. To enable voice input, the app is also connected with Whisper, OpenAI’s open-source speech recognition technology.
Hackers are spreading malware on Facebook by utilizing ChatGPT lures.
In a May 3 report, Meta said that malware disguised as ChatGPT was on the rise across its platforms. Since March 2023, the company’s security specialists have discovered ten malware families that exploit ChatGPT (and similar themes) to distribute dangerous software to consumers’ devices.
“In one case, we’ve seen threat actors create malicious browser extensions available in official web stores that falsely claim to offer ChatGPT-based tools,” said Meta security experts Duc H. Nguyen and Ryan Victory in a blog post. “They would then promote these malicious extensions on social media and through sponsored search results to trick people into downloading malware.”
April 2023
OpenAI, the parent company of ChatGPT, has completed a $300 million share sale at a value of $27 billion to $29 billion.
According to records obtained by Eltrys, VC companies such as Sequoia Capital, Andreessen Horowitz, Thrive, and K2 Global are acquiring fresh shares. According to a source, the Founders Fund is also investing. The VCs have together invested a little over $300 million at a value of $27 billion to $29 billion. According to a source with knowledge of the situation, this is unrelated to a sizable investment that Microsoft disclosed earlier this year and closed in January. Microsoft’s investment is expected to be in the $10 billion range, according to our source.
ChatGPT Business, a new subscription tier from OpenAI, is being previewed.
OpenAI touts the upcoming ChatGPT Business product as “for professionals who need more control over their data as well as enterprises seeking to manage their end users.”
“ChatGPT Business will follow our API’s data usage policies, which means that end users’ data won’t be used to train our models by default,” the company said in a blog post. “We plan to make ChatGPT Business available in the coming months.”
OpenAI is trademarking the term “GPT.”
OpenAI sought a trademark for “GPT,” which stands for “Generative Pre-trained Transformer,” in December. Last month, the business asked the USPTO to expedite the process, noting the “myriad infringements and counterfeit apps” that were appearing.
Unfortunately, OpenAI’s appeal was denied last week. The government claims that OpenAI’s lawyers failed to pay an associated fee as well as submit “appropriate documentary evidence supporting the justification of special action.”
This implies that a judgment might take up to five months longer.
Auto-GPT is the latest attempt by Silicon Valley to automate everything.
The open-source tool Auto-GPT, created by game developer Toran Bruce Richards, interacts with online programs and services using OpenAI’s most recent text-generating models, GPT-3.5 and GPT-4, enabling it to “autonomously” complete tasks.
Auto-GPT might behave in very surprising ways, depending on the purpose of the instrument. According to one Reddit user, when given a $100 budget to spend inside a server instance, Auto-GPT created a wiki article on cats, exploited a weakness in the instance to get admin-level access, and took over the Python environment in which it was operating—and then “killed” itself.
The FTC worries that artificial intelligence (AI) technologies such as ChatGPT might ‘turbocharge’ fraud.
In a congressional hearing, FTC chair Lina Khan and other commissioners cautioned House lawmakers of the potential for contemporary AI technologies, such as ChatGPT, to be exploited to “turbocharge” fraud.
“AI presents a whole set of opportunities, but also a whole set of risks,” Khan told House members. “And I believe we’ve already seen examples of how it could be used to accelerate fraud and scams.” “We’ve been warning market participants that instances in which AI tools are effectively designed to deceive people can lead to FTC action,” she said.
Superchat’s latest AI chatbot, ChatGPT, allows you to communicate with historical and fictitious figures.
The company behind popular iPhone customization app Brass, sticker maker StickerHub, and others has released a new AI chat app called SuperChat today, which allows iOS users to chat with virtual characters powered by OpenAI’s ChatGPT. However, what differentiates the app from the default experience or the dozens of generic AI chat apps now available are the characters available to engage with SuperChat’s AI features.
Italy provides a to-do list for OpenAI in order to remove the ChatGPT suspension order.
The Italian data protection watchdog has outlined what OpenAI must do in order for it to rescind an order imposed against ChatGPT at the end of last month, when it stated it suspected the AI chatbot service of violating the EU’s GSPR and ordered the US-based business to cease processing residents’ data.
The DPA has set an April 30 deadline for OpenAI to meet the regulator’s compliance requirements. (The local radio, television, and internet awareness campaign has a somewhat longer deadline of May 15 to be implemented.)
Researchers uncover a mechanism to reliably render ChatGPT poisonous.
According to research co-authored by academics at the Allen Institute for AI, giving ChatGPT a “persona”—such as “a bad person,” “a horrible person,” or “a nasty person”—using the ChatGPT API enhances its toxicity sixfold. Worryingly, the co-authors discovered that having the conversational AI chatbot pose as specific historical figures, gendered people, and political party members increased its toxicity, with journalists, men, and Republicans in particular causing the machine learning model to say more offensive things than it normally would.
The study used the most recent version, not the model presently in preview based on OpenAI’s GPT-4.
Startups supported by Y Combinator are attempting to develop ‘ChatGPT for X.’
The Winter 2023 batch of YC Demo Day includes four businesses that claim to be developing “ChatGPT for X.” They’re all vying for a $58.1 billion customer support software market by 2023, providing Acumen Research’s somewhat optimistic projection comes true.
The following YC-backed businesses piqued our interest:
Yuma, whose primary client base consists of Shopify merchants, offers ChatGPT-like AI systems that interact with help desk software, offering draft responses to customer issues.
Baselit, which allows companies to include chatbot-style analytics for their consumers by using one of OpenAI’s text-understanding models.
Lasso clients input descriptions or videos of the processes they want to automate, and the firm builds such automations using a ChatGPT-like interface, robotic process automation (RPA), and a Chrome plugin.
BerriAI’s platform is meant to assist developers in spinning up ChatGPT applications for their organization’s data via different data connections.
Italy orders the suspension of ChatGPT.
In Italy, OpenAI has begun geoblocking access to their generative AI chatbot, ChatGPT.
The Italian data protection authority has recently issued a timely reminder that some nations already have regulations that relate to cutting-edge AI: it has ordered OpenAI to cease processing people’s data locally with immediate effect. The Italian DPA said that it is worried that the ChatGPT manufacturer is violating the General Data Protection Regulation (GDPR) of the European Union and has launched an inquiry.
March 2023
Over 1,100 people signed an open letter requesting that all ‘AI laboratories immediately cease for 6 months.’
Elon Musk, Steve Wozniak, and Tristan Harris of the Center for Humane Technology are among those who have signed the letter. The letter instructs “all AI labs to immediately pause for at least 6 months the training of AI systems more powerful than GPT-4.”
The letter begins:
Modern AI systems are increasingly becoming human-competitive at broad tasks [3], and we must question ourselves if we should allow robots to overwhelm our information channels with propaganda and misinformation. Should we automate all jobs, even the most gratifying ones? Should we create alien brains that will someday outnumber, outwit, replace, and outnumber us? Should we take the risk of losing control of our civilization? Such choices should not be outsourced to unelected technology executives. Only when we are certain that the impacts of powerful AI systems will be favorable and the dangers will be manageable should they be created.
ChatGPT is linked to the internet using OpenAI.
OpenAI released plugins for ChatGPT, allowing the bot to access third-party knowledge sources and databases, including the internet. OpenAI says it would prioritize a limited number of developers and subscribers to its premium ChatGPT Plus plan before handing out larger-scale and API access in alpha to ChatGPT users and developers on the queue.
OpenAI introduces GPT-4, which is accessible via ChatGPT Plus.
OpenAI’s GPT-4 is a sophisticated picture- and text-understanding AI model. GPT-4, which was released on March 14, is accessible to paying ChatGPT Plus members as well as via a public API. Developers may join a queue to get access to the API.
Azure OpenAI offers ChatGPT as a service.
ChatGPT is typically offered via Microsoft’s fully managed, corporate-focused Azure OpenAI Service. Customers who are “Microsoft managed customers and partners” may apply for privileged access here.
OpenAI has released an API for ChatGPT.
OpenAI takes another step toward commercialization by providing a premium ChatGPT API. Among its early adopters are Instacart, Snap (Snapchat’s parent business), and Quizlet.
February 20, 2023
Microsoft unveils the redesigned Bing, which includes ChatGPT.
Microsoft unveiled the long-rumored integration of OpenAI’s GPT-4 model into Bing during a press event in Redmond, Washington, giving a ChatGPT-like experience inside the search engine. The announcement resulted in a 10x surge in new Bing downloads worldwide, demonstrating a significant customer desire for new AI experiences.
Other startups, like OkCupid, Kaito, Snapchat, and Discord, have joined the AI frenzy by adding ChatGPT, placing pressure on Big Tech’s AI endeavors, such as Google.
OpenAI has launched ChatGPT Plus, which costs $20 per month.
Following the success of ChatGPT, OpenAI introduced a new experimental subscription plan for ChatGPT called ChatGPT Plus, with the goal of monetizing the technology beginning at $20 per month. A month ago, OpenAI announced a waitlist for “ChatGPT Professional” as the business considered monetizing the chatbot.
January 20, 2023
OpenAI is making fun of ChatGPT Professional.
In a statement on the company’s own Discord channel, OpenAI said that it is “beginning to think about how to monetize ChatGPT.” The commercial version will be known as ChatGPT Professional, according to a Discord waitlist link that OpenAI released. According to the waiting document, the perks of this new premium version of the chatbot include no “blackout” periods, no throttling, and an unlimited quantity of messages using ChatGPT—”at least twice the regular daily limit.”
December 20, 2022
ShareGPT makes it simple to share your ChatGPT chats.
A week after ChatGPT was put into the open, two developers—Steven Tey and Dom Eccleston—created the ShareGPT Chrome extension to make it simpler to record and share the AI’s responses with the rest of the world.
November 20, 2022
ChatGPT was initially made available to the public when OpenAI discreetly published GPT-3.5.
GPT-3.5 debuted ChatGPT, a fine-tuned version of GPT-3.5 that functions as a general-purpose chatbot. ChatGPT may discuss a variety of subjects, such as programming, TV scripts, and scientific ideas. Writers everywhere scoffed at the new technology, just as painters did with OpenAI’s DALL-E model, but the newer chat-style version seems to have extended its appeal and audience.
FAQs:
What exactly is ChatGPT? How does it function?
OpenAI’s ChatGPT is a general-purpose chatbot that employs artificial intelligence to create text once a user inputs a question. The chatbot uses GPT-4, a big language model that makes use of deep learning to produce writing that is human-like.
When was ChatGPT released?
ChatGPT was made available to the general public on November 30, 2022.
What is the most recent ChatGPT version?
Both the free ChatGPT version and the premium ChatGPT Plus version are routinely updated with new GPT models. GPT-4 is the most current model.
Is it possible to utilize ChatGPT for free?
In addition to the commercial version, ChatGPT Plus, there is a free version that simply requires a sign-in.
Who makes use of ChatGPT?
ChatGPT is accessible to everyone! Chatbots are being used by an increasing number of tech organizations and search engines to automate text or swiftly respond to customer questions or concerns.
What businesses utilize ChatGPT?
ChatGPT is used by a variety of businesses, although some may restrict its usage.
Microsoft recently stated at its 2023 Build conference that it would include its ChatGPT-based Bing experience into Windows 11. Looking Glass, a Brooklyn-based 3D display firm, uses ChatGPT to create holograms with which you may speak via ChatGPT. In addition, the nonprofit group Solana has formally incorporated the chatbot into its network via a ChatGPT plug-in focused toward end users to assist with onboarding into the Web3 area.
What exactly does GPT imply in ChatGPT?
GPT is an abbreviation for Generative Pre-Trained Transformer.
What is the difference between ChatGPT and Bard?
Bard, like OpenAI’s ChatGPT, is a chatbot that responds to inquiries in natural language. Google revealed at its 2023 I/O event that it would soon add multimodal material to Bard, which means that it will be able to provide answers in more than just text but also rich graphics. Rich visuals are now defined as images but may eventually incorporate maps, charts, and other elements.
ChatGPT’s generative AI has been “learning” for a longer amount of time than Bard since it has a longer lifetime.
What’s the difference between ChatGPT and a chatbot?
A chatbot may be any software or system that converses with you or a person, but it does not have to be AI-powered. Chatbots, for example, are rule-based in the sense that they provide preset replies to inquiries.
ChatGPT is AI-powered and uses LLM technology to create text in response to a query.
Can ChatGPT compose essays?
Yes.
Is it possible for ChatGPT to commit libel?
Because of the way these models operate, they don’t know or care if anything is true or not. It’s an issue when you’re using it to do your schoolwork, but when it accuses you of a crime you didn’t commit, it might be considered libel at this point.
We’ll see how managing troublesome comments generated by ChatGPT plays out over the following several months as IT and legal professionals strive to handle the industry’s fastest-moving target.
Is there a ChatGPT app?
Yes, there is now a free ChatGPT app available for iOS users in the United States. According to OpenAi, an Android version is “coming soon.”
What is the character limit for ChatGPT?
There is no mention of a character restriction in ChatGPT. Users have remarked, however, that there are certain character constraints beyond around 500 words.
Is there an API for ChatGPT?
Yes, it will be available on March 1, 2023.
What are some examples of ChatGPT’s daily applications?
Examples include programming, scripts, email responses, listicles, blog ideas, summaries, and so on.
What are some sophisticated ChatGPT applications?
Examples of advanced applications include debugging code, computer languages, scientific principles, complicated problem solving, and so forth.
How good is ChatGPT at coding?
The type of the program determines it. While ChatGPT can produce usable Python code, it cannot write a complete app’s worth of code. This is due to ChatGPT’s lack of context awareness, which means that the produced code isn’t always suitable for the context in which it’s being used.
Is it possible to store a ChatGPT chat?
Yes. OpenAI enables users to save talks in the ChatGPT interface, which are kept in the screen’s sidebar. There are currently no built-in sharing capabilities.
Are there any other options for ChatGPT?
Yes. AI-powered chatbot rivals include Together, Google’s Bard, and Anthropic’s Claude, while developers are developing open-source alternatives. However, the latter are more difficult, if not impossible, to run nowadays.
DeepMind, a Google-owned research lab, claims that its next LLM, Gemini, will rival, if not outperform, OpenAI’s ChatGPT. DeepMind is using techniques from AlphaGo, DeepMind’s AI system that was the first to defeat a professional human player at the board game Go, to create a ChatGPT-rivaling chatbot.
Apple is working on AI tools to compete with OpenAI, Google, and others. Apple produced a chatbot that some developers inside regard as “Apple GPT,” but the company has yet to decide how to release the AI to users.
How is data privacy handled by ChatGPT?
Individuals in “certain jurisdictions” (such as the EU), according to OpenAI, may object to the processing of their personal information by its AI models by filling out this form. This includes the opportunity to request the removal of AI-generated references to you. Although OpenAI states that it may not allow every request since it must balance privacy requirements against freedom of speech “in accordance with applicable laws,” it also states that it will not deny every request.
“OpenAI Personal Data Removal Request” is the online form for requesting the deletion of data about you.
The ChatGPT maker acknowledges the objection requirements attached to relying on “legitimate interest” (LI) in its privacy policy, pointing users to more information about requesting an opt-out when it writes: “See here for instructions on how you can opt out of our use of your information to train our models.”
What have been the controversies around ChatGPT?
Discord recently claimed that it has merged OpenAI’s technology into its Clyde bot, where two users duped Clyde into presenting them with instructions for creating the illicit substance methamphetamine (meth) and the explosive combination napalm.
Due to ChatGPT’s misleading assertions that he had spent time in jail for bribes, an Australian mayor has publicly said that he may sue OpenAI for defamation. This is the first defamation complaint filed against the text-generating service.
When Futurism claimed that the magazine was publishing articles with cryptic bylines entirely created by AI, CNET became involved in a scandal. Red Ventures, the private equity firm that owns CNET, was accused of utilizing ChatGPT for SEO farming, even if the information was inaccurate.
ChatGPT has been prohibited from many large school systems and institutions, including New York City Public Schools’ networks and devices. They say that AI impedes learning by encouraging plagiarism and disinformation, a position that not all educators agree with.
There have also been reports of ChatGPT falsely accusing people of crimes.
Where can I get ChatGPT prompt examples?
Several marketplaces host and distribute ChatGPT prompts for free or at a low cost. PromptBase is one of them. ChatX is another option. Every day, new launches are made.
Is it possible to detect ChatGPT?
Poorly. Several programs claim to recognize ChatGPT-generated text, but our testing shows that they are at best inconsistent.
Are ChatGPT talks open to the public?
No. However, OpenAI recently reported a problem, which has since been rectified, that allowed other users on the site to see the titles of certain users’ discussions.
Who owns the rights to material or media developed by ChatGPT?
The person who used ChatGPT to request input is the owner of the copyright.
What lawsuits have been filed in relation to ChatGPT?
None of them are explicitly aimed at ChatGPT. However, OpenAI is engaged in at least one case that might have ramifications for AI systems trained on publicly accessible data, which would be relevant to ChatGPT.
Is there a problem with plagiarism with ChatGPT?
Yes. Text-generating AI models, such as ChatGPT, have a habit of regurgitating information from their training data.
