From Nonprofit to $29 Billion Valuation – The Promise and Danger of OpenAI

The research lab behind the viral ChatGPT chatbot, OpenAI, is in talks to sell existing shares in a tender offer that would value the company at around $29 billion, according to people familiar with the matter, reported the Wall Street Journal on January 5.

This would make OpenAI, which started life as a nonprofit and generates virtually no revenue, one of the most valuable U.S. startups.

Chatbots have captured the imagination of users and investors alike. They provide fast, succinct, and outwardly accurate responses to detailed questions well beyond the capabilities of Alexa, Siri and Google search. Those queries might include providing the foundation for writing an article, like the one you are currently reading, in the style of IPWatchdog. Spoiler alert: a chatbot was used for illustration purposes only (see below).

ChatGPT, a beta product of the company OpenAI, has generated the bulk of this interest. Funded by an all-star list of investors, the “AI research and deployment company” began life as an Elon Musk-founded nonprofit. It has since been spun out into a for-profit whose valuation is soaring.

OpenAI’s stated mission is to ensure that “artificial general intelligence (AGI)—by which we mean highly autonomous systems that outperform humans at most economically valuable work—benefits all of humanity.”

OpenAI debuted its ChatGPT beta in late November, attaining one million users in five days. It took Netflix 41 months, Facebook 10 months, and Instagram 2.5 months to gain a similar following. According to a Reuters source, the company’s most recent valuation is a reflection of a secondary share sale. Microsoft, whose Bing search engine has sputtered, has put in $1 billion. Projected 2024 OpenAI revenues are $1 billion.

Professor Dr. Alexander Wurzer, Director of IP Management Education at the Center for International Intellectual Property Studies (CEIPI) in Strasbourg France, recently suggested that “AI-assisted framing does an excellent job in this environment of making non-IP audiences aware of IP’s relevance to their own professional work.” Professor Wurzer cites Franklin Graves’ timely article in IPWatchdog, “A Busy 2022 for AI and IP Promises Even More in 2023” as reason for optimism.

Professor Wurzer makes a good point: there is an unusually high disconnect regarding the use and purpose of IP rights among most executives and many creators, investors and consumers. AGIs are a potentially powerful tool for providing the content and data people need when they need it. Despite the attention and investment heaped on them, experimental chatbots like ChatGPT and Google’s LaMDA are not without their technical and legal weaknesses

IP professionals, especially, should be wary of chatbots’ potential to infringe IP rights, such as copyrights, as well as the accuracy of their confident-looking outputs. Users are easily seduced by the speed and apparent ability of natural language searches to produce neatly summarized, nuanced responses to complex queries. They are often inaccurate.

Solution or Problem?

“ChatGPT, as currently conceived, is a parlor trick,” Gartner vice president Bern Elliot told CNBC. “It’s something that isn’t actually itself going to solve what people need, unless what they need is sort of a distraction.”

ChatGPT (GPT stands for Generative Pre-trained Transformer) was trained on an enormous amount of text data. It learned to recognize patterns that enable it to produce its own text mimicking various writing styles, said Elliot. OpenAI doesn’t reveal what precise data was used for training ChatGPT.

Elliot said that for now ChatGPT is “more of a way for OpenAI to gain publicity and to show what’s possible for large language models, as opposed to a useful piece of software for businesses to incorporate.”

While ChatGPT is effectively no-cost in its current form, OpenAI sells access to its underlying language and related AI models for businesses to use. A pricing model for the public is likely forthcoming. There has been no discussion about sponsorship or advertising, but that will come sooner and perhaps more subtly than we think.

Trained Responses

ChatGPT is a bot trained to generate human-like responses to user inputs, writes Bloomberg Businessweek. It can produce basic software code, rudimentary financial analysis, amusing poems and songs, natural language summaries of scientific concepts and reflective essays (don’t expect it to herald the end of the term paper). It learns as it goes.

How AGIs gather data, what content they are using, their potential biases (racial and otherwise) have yet to be revealed. Users will need to discern what is (1) accurate and (2) useful in the context of their needs.

Chatbots can produce outwardly precise responses in seconds. They can be highly satisfying and at the same time somewhat frustrating or utterly incorrect. These post-Google search algorithms play to readers’ shorter attention spans and increasing time constraints. ChatGPT gets the presentation right in ways that Internet search engines like Bing and Google are unable to, but frequently at cost to accuracy.

Attention span deficit has become an issue not only for the less sophisticated, unaccustomed to serious research and reams of data, but for virtually everyone using the Internet and social media. Bombarded with information and images and easily distracted, readers need help getting to what they need faster. Why wade through a limited set of articles on Google search to gather the information required to write a paper when you can have a bot compose it? Attention deficit is much less a matter of intelligence than a function of time and patience. Natural language AI bots help already inundated people cut to the chase – or, at least, appear to. Siri and Alexa are sure to want to get in on the action.

Viral Sensation

ChatGPT debuted in late November and has quickly turned into a viral sensation, with people tweeting questions, such as “Are NFTs dead,” and requests like, “Tell a funny joke about the tax risks of international remote work.” They include a screenshot of ChatGPT’s response, which often — but not always — makes sense.

The tech was developed by San Francisco-based OpenAI, – reports CNBC, a research company led by Sam Altman and backed by Microsoft, LinkedIn co-founder Reid Hoffman, co-founder of LinkedIn and a Greylock Partner, and Khosla Ventures. Other investors include Elon Musk, Sequoia Capital, one of the most successful technology investors, Tiger Global Management, Bedrock Capital, Combinator Y, Peter Thiel, PayPal founder and first Facebook investor, and Silicon Valley private equity icon, Andreessen Horowitz.

ChatGPT automatically generates text based on written prompts in a fashion that is much more advanced and creative than the chatbots of Silicon Valley’s past. (It should be an easy transition to voice prompts.) Tech investors are pouring billions of dollars into startups specializing in the field of generative AI, which refers to the ability of computers to automatically create text, videos, photos and other media using cutting-edge machine learning technologies.

There is good reason to believe that ChatGPT is a ‘code red’ for Google Search. That it will change the way we consider search, just as TikTok has for GenZ. Google has been working on its own chatbot, LaMDA, but is currently less eager to go public with it. It can afford to be patient. Why disrupt its current lucrative business model, which generated $54.5 billion in advertising revenue in the third quarter of 2022? It is important to remember that when Google’s search algorithm debuted in 1998, query responses were largely objective. Today, responses are heavily influenced by advertisers’ agendas, as well as Google’s attitude toward a host of topics, such as IP rights. (On January 9, 1998 Sergei Brin filed a U.S. patent application for the PageRank algorithm, which issued on September 4, 2001.)

Follow the Money

The challenges of intelligent systems for search include an initial learning lag. Similar to AI systems that guide autonomous vehicles, they are very impressive, but they learn from their mistakes and are not ready for primetime. This is painfully clear when an AV system fails. With a chatbot, responses delivered in crisp, declarative sentences may appear definitive. Often, they are not. That may be acceptable to some, but the fact is, chatbot responses are frequently less accurate than they appear.

To understand how something happens, follow the money. Considering the investors mentioned above, the core code to OpenAI may continue to be open to developers as the company evolves, but don’t expect ChatGPT to remain complimentary for long—oor free of influential users and sponsors.

Outsmarting the Bots

The term paper and essay as we know it is not dead. Educators will need to be wiser than the smart systems that serve to generate them about detecting “artificial” output and plagiarism. Perhaps teachers will need bots to detect the accuracy of chatbot responses and the identity of those making the queries? Currently, Grammarly’s Google’s Plagiarism Checker offers some help. No doubt the definition of plagiarism will evolve. A Wall Street Journal reporter attempted to write her Advanced Placement essay with ChatGPT and passed, although the teacher did identify weaknesses in her, or rather “its”, work.

Will chatbots marginalize human thinking or inspire people to higher levels of thought and analysis? In the end, the bots are a product of us, and we are constantly adapting and so will they. I think they will ultimately inspire us to deeper thinking.

Open AI admits it is an experimental bot that “sometimes writes plausible-sounding but incorrect or nonsensical answers.” Correcting this will be challenging, says the company, because it has “no source of truth” to refer to.

Tools that provide reliable data, if not perfect answers, are wonderful if we use them wisely. We (humans) have to learn what chatbot responses are and are not, and how to use them most effectively. For now, chatbots are in a position to inspire fast, succinct and generally accurate responses to basic questions about topics like IP rights and how they are used. We need to keep an eye on these responses, not only for accuracy, but for bias, given chatbots’ capital and revenue sources, as well as who programs them.

It is ironic that the very AI search for better IP understanding is likely a violation of someone’s copyright (but that is another discussion for another Intangible Investor).

Preview of Progress

There will be competition on the top end. The search giants and tech funders battling each other for position. But don’t be surprised if businesses and individual users will require bots to check on other bots’ accuracy, in a kind of AI “War of the Worlds” (or is that “War of the Words?”). Will there be a battle of the bots when it comes to reliability and accuracy? It will be important that busy users care.

I once heard someone comment that Google is a killer algorithm away from obscurity. The tongue-in-cheek remark is a point well taken. Google will not give up its title to 92.2% of global search without a fight. It has its own chatbot in the wings, which it is refining. It is reluctant to move too quickly less it threaten its own business model. Still, the company is concerned that it could very well go the way of once-successful businesses like Webcrawler, Lycos and AskJeeves.

“Although ChatGPT still has plenty of room for improvement, its release led Google’s management to declare a ‘code red,’ reports The New York Times. “For Google, this was akin to pulling the fire alarm. Some fear the company may be approaching a moment that the biggest Silicon Valley outfits dread — the arrival of an enormous technological change that could upend the business.”

A recent Tweet from OpenAI co-founder Sam Altman may offer Google shareholders and executives some comfort:

“ChatGPT is incredibly limited, but good enough at some things to create a misleading impression of greatness. it’s a mistake to be relying on it for anything important right now. it’s a preview of progress; we have lots of work to do on robustness and truthfulness.” 

Example 1: Safer Than Useful

The following is the ChatGPT response to the question, “What is the best way to learn about the purpose of intellectual property rights?” The response is safe and generic, offering less specific help than a Google search.

There are a number of ways that you can learn about the purpose of intellectual property (IP) rights, including:

1. Researching online: There are many online resources that provide information about the purpose of IP rights and how they are used to protect and promote innovation and creativity. You can find articles, guides, and other resources on websites such as the World Intellectual Property Organization (WIPO) or the United States Patent and Trademark Office (USPTO).
2. Consulting with a qualified attorney or other expert: If you have specific questions or concerns about the purpose and use of IP rights, you can consult with a qualified attorney or other expert who is familiar with the relevant laws and regulations. They will be able to provide you with more detailed and accurate information and guidance.
3. Enrolling in a course or program: There are many educational programs and courses that cover the basics of IP law and the purpose of IP rights. These programs can be found at universities and other educational institutions, or online through websites like Coursera or edX.
4. Reading relevant books or articles: There are many books and articles that provide detailed information about the purpose and use of IP rights. You can find these resources at your local library or online through websites like Amazon or Google Books.

Overall, the best way to learn about the purpose of IP rights will depend on your specific learning style and needs. It is generally a good idea to research and learn from multiple sources in order to get a well-rounded understanding of the topic.

Example 2: Surprisingly Nuanced

This response to “What do students need to know about intellectual property rights?” is more specific and surprisingly accurate. With chatbots, like traditional search engines, a successful response is predicated on not just what you ask about but how you ask it.

There are a number of things that students should know about intellectual property (IP) rights:

1. IP rights protect and promote innovation and creativity: IP rights, such as patents, trademarks, and copyrights, are designed to encourage innovation and creativity by giving creators and inventors exclusive rights to their creations and inventions. This allows them to benefit financially from their work and encourages them to continue creating and innovating.
2. IP rights are limited in scope and duration: IP rights are not unlimited and are only granted for a specific period of time. This means that after the term of protection expires, the work or invention becomes part of the public domain and can be used freely by anyone.
3. IP rights are subject to certain exceptions and limitations: While IP rights give creators and inventors certain exclusive rights, there are also exceptions and limitations to these rights. For example, copyright law generally allows for the use of a work for the purpose of criticism, commentary, news reporting, teaching, scholarship, or research, without the need for permission from the copyright holder.
4. IP rights are enforceable: If someone infringes on your IP rights, you have the right to take legal action to protect your rights and seek damages.
5. IP rights can be bought, sold, and licensed: IP rights can be bought and sold, just like any other property, and can also be licensed to others for a fee. This allows creators and inventors to monetize their work and allows others to use it in a way that is consistent with the terms of the license.
6. As a student, you may need to obtain permission to use someone else’s work: If you are using someone else’s work as part of your studies or research, you may need to obtain permission from the owner of the work. This is particularly important if you are planning to publish your work or distribute it to others.

Overall, it is important for students to understand the purpose and limitations of IP rights in order to respect the rights of creators and inventors and to ensure that innovation and creativity can thrive. It is also important for students to be aware of the rules and regulations surrounding the use of others’ work in order to avoid infringing on their IP rights.