Potential Risks of ChatGPT and Other Generative AI | Baker Hostetler

AI For Business

“Shall we play a game?” Harmless words “uttered” by Matthew Broderick’s computer in John Badham’s sci-fi techno thriller war games It stunned the audience at the time. A computer that can “talk”, “think” and “converse”!? This was the height of science fiction. Recent releases of generative artificial intelligence (AI) tools, specifically in the form of ChatGPT and other predictive natural language processing (NLP) algorithms, have made science fiction real again.

Companies from Microsoft to Google, Instacart to Kayak are beginning to embed and build on this technology originally developed by OpenAI. These tools are very beneficial to your business, but they also come with risks.

Before we dive into how generative AI can help brands and businesses, let’s first dig into the layers to understand what ChatGPT is at a basic level.

Abbreviation for ChatGPT Hahat G.generative P.retrained T.runner. Let’s break it down:

  • chat Refers to an interface that allows interaction with a model using natural language prompts.
  • driving force Refers to a category of AI models that generate new outputs based on given inputs. In practice, this means that the “input” of the user’s query produces the “output” of the text, image and voice answers.
  • pre-trained It refers to the fact that the model has already been trained on a huge dataset and taught to predict the next word in a given sequence.
  • transformer Refers to the underlying neural network (machine learning algorithm) architecture of ChatGPT. It is this architecture that enables computers to process natural language.

Importantly, ChatGPT and other generative AIs are not omnipotent. They cannot think, understand, or feel. They are simply pieces of software (lines of computer code) programmed to generate natural language responses in response to text or image prompts. They work by predicting the next word for a given text string based on patterns “learned” from the data they were trained on.

In late March, OpenAI released an API (application programming interface) to allow businesses of their choice to incorporate AI technology into their websites and apps via plugins. By using these plugins, brands can harness the power of his ChatGPT to help consumers book trips, book restaurants, and make curated product recommendations. became.

Currently, the ChatGPT plugin is assigned relatively basic functionality, essentially providing a high-level search tool in the form of an interactive chatbot. For example, see the video published by Expedia. twitter Here’s how that ChatGPT plugin works.

Brands are also beginning to use generative AI tools to help:

  • Coding – generation and construction of source code and analysis of errors in code
  • Content creation – generate blog posts, social media posts, targeted email campaigns, video scripts
  • Data Analysis – Analyze large datasets and consolidate information into easily digestible bullet points
  • Market Research – Create a list of key players in any industry along with their products and services
  • Product descriptions – Generate batches of descriptions for e-commerce sites with frequently updated product catalogs.
  • Search engine optimization (SEO) – generate copy with keywords and meta descriptions that search engines can search for when ranking your page

ChatGPT and other generative AI tools carry risks and should be used sparingly and cautiously in business.

First, generative AI is as good and accurate as the information it was trained on. A huge amount of data (approximately 570 GB) was used to train the ChatGPT model, but the output is “inaccurate, untrue, and misleading.” [1] It also “occasionally generates incorrect answers… Harmful instructions and biased content”.[2] Furthermore, ChatGPT is “confident and makes a fair percentage of mistakes” and the tool is “good enough in some ways to give the misleading impression that it’s good.” ”.[3] In other words, don’t rush to replace human customer service teams with generative AI bots. Make sure humans are constantly observing AI-generated responses.

Second, using output from generative AI in published work can carry copyright risks. Due to how the ChatGPT model was trained, the responsive output may contain elements of copyrighted material. Brands that use AI-generated materials in popular ways, such as advertisements or social media posts, may run the risk of violating the rights of others for which the brand may be held liable. So you can use these tools as a starting point for your ideas, but don’t use them yet to produce the final deliverable.

Moreover, under current US intellectual property law, copyright protection exists only where there is human selection and placement of information. There is no copyright protection for material created by non-humans.[4] However, recent guidance states that for works containing AI-generated material, the Copyright Office will require that “either the AI ​​contribution is the result of mechanical duplication” or that the author “has given a visible form”. It suggests to consider whether it is “the original mental concept of the author himself”. As such, it is unclear whether material produced by generative AI without significant human input can receive copyright protection.

Third, there are potential confidentiality issues to consider. Under US law, to receive trade secret protection, a company must use reasonable efforts to conceal that information from the public. Allowing a third party, such as OpenAI, to access otherwise protected information without entering into a non-disclosure agreement constitutes a “public disclosure” and the protection of the disclosed information It is in danger of being lost.[5]

In addition to the above, there are privacy, security, ethical, and other legal risks to consider when using this technology or allowing third parties to access internal customer records and business data. . For more information on these issues, read our sister blogs here, here, and here.

In the words of Stan Lee, “With great power comes great responsibility.” New technologies are revolutionary and help improve productivity and customer service. So move forward. But tread lightly. Also, before implementing any of these tools, consult with legal counsel to assess what legal risks their use may pose.

[1] OpenAI, “What is ChatGPT?”, https://help.openai.com/en/articles/6783457-chatgpt-general-faq (last visited 10 April 2023).

[2] identification.

[3] Berber Jin and Miles Kruppa, “The Story Behind ChatGPT Creator OpenAI” wall street journal (2:57 PM ET Dec. 18, 2022), https://www.wsj.com/articles/chatgpt-creator-openai-pushes-new-strategy-to-gain-artificial-intelligence-edge-11671378475 , and Karen Hao, “What is ChatGPT? What You Need to Know About AI Chatbots” wall street journal (22 March 2023), https://www.wsj.com/articles/chatgpt-ai-chatbot-app-explained-11675865177.

[4] United States Copyright Office, Overview of U.S. Copyright Office Practices § 313.2 (3d ed. 2021). Discover Urantia. v. Kristen Mahera114 F.3d 955, 957-59 (9th Cir. 1997) (“Some element of human creativity [b]because “it’s not God’s creation that copyright law was meant to protect”).

[5] According to OpenAI, information provided to the service may be used to “provide and maintain the service,” and OpenAI may “review and improve conversations.” [its] system. https://openai.com/policies/terms-of-use and https://help.openai.com/en/articles/6783457-chatgpt-general-faq.

[View source.]

Source link

Leave a Reply

Your email address will not be published. Required fields are marked *