Study with Columbia University on Spinocerebellar Ataxias

What is OpenAI’s ChatGPT?

Created by San Francisco-based tech lab OpenAI, ChatGPT is a generative AI software application that uses a machine learning technique called reinforcement learning from human feedback (RLHF) to emulate human-written conversations based on a large range of user prompts. This kind of software is better known as an AI chatbot.

ChatGPT learns language by training on texts gleaned from across the internet, including online encyclopedias, books, academic journals, news sites and blogs. Based on this training, the AI chatbot generates text by making predictions about which words (or tokens) can be strung together to produce the most suitable response.

More than a million people engaged with ChatGPT within the first week of its launch for free public testing on November 30, 2022. Many were in awe of the chatbot’s seemingly natural language capabilities, not only in terms of understanding questions, but also because of its human-like responses. Users felt as if they were having a conversation with a human.

Besides being an excellent conversation partner, ChatGPT can write engaging short stories, develop catchy marketing materials, solve complicated math problems, and even create code in various programming languages.

Based on this success, OpenAI created a more powerful version of the ChatGPT system called GPT-4, which was released in March 2023. This iteration of ChatGPT can accept visual inputs, is much more precise and can display a higher level of expertise in various subjects. Because of this, GPT-4 can describe images in vivid detail and ace standardized tests.

Unlike its predecessor, GPT-4 doesn’t have any time limits on what information it can access; however, AI researcher and professor Dr. Oren Etzioni has said that the chatbot is still terrible at discussing the future and generating new ideas. It also hasn’t lost its tendency to deliver incorrect information with too high a degree of confidence.

Further improving on its product, in May 2024 OpenAI launched Chat GPT-4o, with the o standing for omni. OpenAI describes GPT-4o as “a step towards much more natural human-computer interaction—it accepts as input any combination of text, audio, image, and video and generates any combination of text, audio, and image outputs.”

This version has done away with the lagging response time afflicting GPT-4. This proves especially helpful for producing immediate translations during conversations between speakers of different languages. It also allows users to interrupt the chatbot to pose a new query to modify responses.

Why is Microsoft investing in OpenAI?

Ascannio / Shutterstock

Since 2019, Microsoft has invested at least US$3 billion in OpenAI to help the small tech firm create its ultra-powerful AI chatbot, as reported by New York Times technology correspondents Cade Metz and Karen Weise.

Microsoft announced in mid-January 2023 that as part of the third phase of its partnership with OpenAI, it will make “a multiyear, multibillion dollar investment.” Although the company hasn’t disclosed the total amount of its latest spend, reports at the time indicated that US$10 billion is on the table. According to a February article from Reuters, OpenAI was recently valued at US$80 billion, meaning Microsoft’s US$10 billion move would be huge. However, as of late 2023 there were rumors that OpenAI has only received a fraction of that purported investment.

How could Microsoft benefit from its investment? It seems the tech giant is hopeful advancements in generative AI may increase revenues for its Azure cloud computing business, as OpenAI officially licensed its technologies to Microsoft in 2020. Indeed, Pitchbook has described the deal as an “unprecedented milestone” for generative AI technology.

The strength of Microsoft’s confidence in OpenAI’s Altman was definitely on display in late November, when it quickly moved him to the payroll of its advanced AI research team after he was fired from OpenAI. Barely a week passed before Altman was back at the helm of OpenAI with major board changes, including the addition of Dee Templeton, Microsoft’s vice president of technology and research partnerships and operations, as a non-voting observer.

What is Elon Musk’s relationship to OpenAI?

DIA TV / Shutterstock

OpenAI was founded in 2015 by Altman, its current CEO, as well as Tesla’s (NASDAQ:TSLA) Elon Musk and other big-name investors, such as venture capitalist Peter Thiel and LinkedIn co-founder Reid Hoffman. Musk left his position on OpenAI’s board of directors in 2018 to focus on Tesla and its pursuit of autonomous vehicle technology.

A few days after ChatGPT became available for public testing, Musk took to X, formerly known as Twitter, to say, “ChatGPT is scary good. We are not far from dangerously strong AI.” That same day, he announced that X had shut the door on OpenAI’s access to its database so it could no longer use it for RLHF training.

His reason: “OpenAI was started as open-source & non-profit. Neither are still true.”

Furthering his feud with OpenAI, Musk filed a lawsuit against the company in March 2024 for an alleged breach of contract. The crux of his complaint was that OpenAI has broken the “founding agreement” made between the founders (Altman, Greg Brockman and himself) that the company would remain a non-profit. Altman and OpenAI have denied there was such an agreement and that Musk was keen on an eventual for-profit structure.

Musk dropped the lawsuit three months later without giving a reason, reported Reuters. The day before he dropped the lawsuit, he reacted to the news that Apple (NASDAQ:AAPL) is partnering with OpenAI to incorporate ChatGPT with Apple devices. On X, Musk declared, “If Apple integrates OpenAI at the OS (operating system)level, then Apple devices will be banned at my companies. That is an unacceptable security violation.” It should be noted that OpenAI has said queries completed on Apple devices will not be stored by OpenAI.

Is ChatGPT revolutionary or hype?

Is ChatGPT a revolutionary technology or just another hyped-up tech fad that will flop, much in the way of Google Glass or the Segway? It may be too early to tell, but as with any new technology, there are plenty of wrinkles to iron out.

One of the most challenging bugs to fix before ChatGPT can be deployed more widely is the chatbot’s propensity to respond with “plausible-sounding but incorrect or nonsensical answers,” admits OpenAI.

Remember, its selection of which words to string together are actually predictions — not as fallible as mere guesses, but still fallible. Even the 4.0 version is “still is not fully reliable (it “hallucinates” facts and makes reasoning errors),” says the company, which emphasizes that users should exercise caution when employing the technology.

Indeed, ChatGPT’s failings can have dangerous real-life consequences. Among other negative applications, the tech can be used to spread misinformation, carry out phishing email scams or write malicious code. What’s more, the AI-based technology is prone to racial and gender-based biases. Not only has this language learning model contributed to the human-like quality of its responses, but it has also picked up on some of humanity’s shortcomings.

“ChatGPT was trained on the collective writing of humans across the world, past and present. This means that the same biases that exist in the data, can also appear in the model,” explains Garling Wu, staff writer for online technology publication MUO, in a September 2023 article. “In fact, users have shown how ChatGPT can give produce some terrible answers, some, for example, that discriminate against women. But that’s just the tip of the iceberg; it can produce answers that are extremely harmful to a range of minority groups.”

On the flip side, an August 2023 study by the University of East Anglia identified a left-wing bias in ChatGPT. Researchers at the school said their work shows that ChatGPT “favors Democrats in the U.S., the Labour Party in the U.K., and president Lula da Silva of the Workers’ Party in Brazil,” according to Forbes.

There’s also the fear among teachers that the technology is leading to an unwelcome rise in academic dishonesty, with students using ChatGPT to write essays or complete their science homework.

“Teachers and school administrators have been scrambling to catch students using the tool to cheat, and they are fretting about the havoc ChatGPT could wreak on their lesson plans,” writes New York Times tech columnist Kevin Roose.

Cybersecurity risks are also a concern for ChatGPT users, and recent events along these lines add validity to Musk’s warning. For one, Mac GPT (the ChatGPT for macOS) was discovered to be breaching Apple’s security rules by storing data as plain text rather than encryption, making it easier for more nefarious apps to access.

Despite these concerns, we’re likely to see new iterations of ChatGPT — hopefully without the aforementioned bugs — as OpenAI has the backing of tech giant Microsoft.

What’s the future of OpenAI and ChatGPT?

The ChatGPT 3.5 platform is free to use, and can be accessed via the web. Those with an iPhone or iPad can also use ChatGPT through an app, and an Android version launched in July 2023. OpenAI also launched a paid subscription, ChatGPT Plus for business use, in August 2023. ChatGPT Plus gives users access to GPT-4, and the newest iteration GPT-4o.

In addition to Microsoft’s use of the ChatGPT technology as part of Copilot, other companies are working with OpenAI to incorporate the technology into their platforms, including Canva, Duolingo (NASDAQ:DUOL), Intercom, Salesforce (NYSE:CRM), Scale, Stripe, and Upwork (NASDAQ:UPWK).

As uptake increases, generative AI technology is replacing humans in the workplace, and will likely continue doing so in a number of fields, from content creation and customer service to transcription and translation services, and even in graphic design and paralegal fields. However, humans are hitting back, as evidenced by recent lawsuits launched against OpenAI and Microsoft. As mentioned, a growing group of prominent authors is suing the creator of ChatGPT and its financial backer for infringing on their copyright by using their books without permission to train the language models behind ChatGPT and other AI-based software.

The New York Times has also taken a stand by taking OpenAI and Microsoft to Manhattan Federal Court.

“Defendants seek to free-ride on the Times’s massive investment in its journalism by using it to build substitutive products without permission or payment,” states the complaint. “There is nothing ‘transformative’ about using the Times’s content without payment to create products that substitute for the Times and steal audiences away from it.”

Scarlett Johansson has also entered the ChatGPT legal minefield after she discovered OpenAI using what she claims is her voice for its chatbot personal assistance voice, Sky. CEO Sam Altman, however, has denied using her voice without permission. “The voice of Sky is not Scarlett Johansson’s, and it was never intended to resemble hers. We cast the voice actor behind Sky’s voice before any outreach to Ms. Johansson,” said Altman in a May 20, 2024 statement. “Out of respect for Ms. Johansson, we have paused using Sky’s voice in our products. We are sorry to Ms. Johansson that we didn’t communicate better.”

What about the long-term goals for OpenAI and ChatGPT? Metz of the New York Times believes the end game is “artificial general intelligence, or AGI — a machine that can do anything the human brain can do.”

In keeping with this end goal, OpenAI made a major move by acquiring an AI creative firm with a deep talent bench, Global Illumination, in mid-August 2023. “Global Illumination is a company that has been leveraging AI to build creative tools, infrastructure, and digital experiences,” states OpenAI on its website.

“The team previously designed and built products early on at Instagram and Facebook and have also made significant contributions at YouTube, Google, Pixar, Riot Games, and other notable companies.”

In November 2023, OpenAI decided to give customers without coding skills the ability to create customized versions of its chatbot and access to large data sets for training. “OpenAI wants people to start innovating using the chatbots and creating special chatbots,” Hod Lipson, an engineering and data science professor at Columbia University, told CNBC.

Chatbot creators will eventually have the ability to share their custom chatbots through OpenAI’s GPT Store. “They’re really trying to create a marketplace, which will allow companies and people to innovate and play around with this incredible form of AI that they’ve just unleashed,” Lipson added.

What is Google’s Gemini?

While ChatGPT has been generating major buzz, it’s definitely not the only chatbot out there.

Notably, Alphabet (NASDAQ:GOOGL) subsidiary Google launched its answer to ChatGPT in March 2023. Originally known as Bard AI, the chatbot is built on Google’s Language Model for Dialogue Applications (or LaMDA). Google CEO Sundar Pichai has described Bard as an “experimental conversational AI service … (that) seeks to combine the breadth of the world’s knowledge with the power, intelligence and creativity of our large language models.”

As with ChatGPT, users can key in a query, request or prompt and it will provide a human-like response. One way in which Google’s chatbot may have had a leg up on the original ChatGPT is that the latter could only use data up to 2021, while the former can access up-to-date information online; this is less relevant now that GPT-4 no longer has this limitation.

However, this ability to access current data hasn’t spared it from ChatGPT’s biggest folly: confidently stating misinformation as fact. The Verge reported that when asked about new discoveries from the James Webb Space Telescope, Google’s chatbot “made a factual error in its very first demo.”

In early in 2024, Google launched the latest iteration of its Bard Advanced AI chatbot with a new name, Gemini AI. The new version is powered by Google’s Gemini Ultra large language model.

Which stocks will benefit the most from AI chatbot technology?

While most companies specializing in generative AI remain in the venture capital stage, there are plenty of AI stocks for those interested in the space. INN’s article 5 Canadian Artificial Intelligence Stocks, ASX AI Stocks: 5 Biggest Companies, and 12 Generative AI Stocks to Watch as ChatGPT Soars includes some examples.

Other than companies directly tied to generative AI technology, which stocks are likely to get a boost from advances?

There are several verticals in the tech industry with indirect exposure to AI chatbot technology, such as semiconductors, network equipment providers, cloud providers, central processing unit manufacturers and internet of things.

Some of the publicly traded companies in these verticals include:

• Graphics processing unit leader Nvidia (NASDAQ:NVDA)
• The world’s largest semiconductor chip manufacturer by revenue, Taiwan Semiconductor Manufacturing Company (NYSE:TSM)
• Computer memory and data storage producer Micron Technology (NASDAQ:MU)
• Digital communications firm Cisco Systems (NASDAQ:CSCO)
• Networking products provider Juniper Networks (NYSE:JNPR)
• Semiconductor producer Marvell Technology Group (NASDAQ:MRVL)
• Cloud-computing Amazon Web Services’ parent company Amazon (NASDAQ:AMZN)
• Bluechip multinational technology company IBM (NYSE:IBM)
• Major semiconductor chip manufacturer Intel (NASDAQ:INTC)

Investors who don’t like to put all their eggs in one basket can check out these 5 Artificial Intelligence ETFs. And if you’re looking for a more general overview of the market, INN has you covered with How to Invest in Artificial Intelligence. You can also take a look back at the market in 2023 with our AI Market 2023 Year-End Review, or read projections for AI this year in our AI Market Forecast: 3 Top Trends that will Affect AI in 2024. Generative AI is also a major theme in the Top 10 Emerging Technologies to Watch in 2024.

When will OpenAI go public?

OpenAI stock is not currently publicly traded and as of early July 2024, there are no plans for an OpenAI IPO on the horizon. For now, investors can gain exposure through related tech companies discussed here.

For example, if Microsoft does take a large position in the company, investors will be able to gain indirect exposure to OpenAI by purchasing Microsoft shares. For those seeking direct exposure, be on the lookout for news of an initial public offering (IPO).

When to expect ChatGPT 5?

OpenAI filed a trademark application for ChatGPT-5 in mid-July 2023, which hinted that the next iteration of the generative AI technology is currently under development. There were rumors the company planned to complete training for ChatGPT-5 by the end of 2023, but this did not materialize.

PC Guide noted in April 2024 that Sam Altman had teased an “amazing new model this year” in a March 2024 interview on the Lex Fridman podcast. More recently, tech writer Suswati Basu shared that OpenAI confirmed in a May 28 blog that a new model is in the works, and she predicts an expected release in late 2024 or early 2025.