Artificial Intelligence, Real Concerns: Navigating AI Legal Landmines

Artificial Intelligence robot hand on computer keyboard

Artificial Intelligence, Real Concerns: Navigating AI Legal Landmines

Artificial Intelligence (AI) is all the rage these days. You can’t be online without hearing about it.  But AI isn’t new.  In fact, Artificial intelligence  revolutionized the way we live and work. If you ever asked Siri to call home or Alexa to give you the weather report, you used AI.

Even your GPS is AI! As someone who used paper maps and spent hours plotting vacation routes pre-computer (yes, I’m that old!), I truly appreciate the convenience AI affords us in our everyday lives.  This article discusses the key legal concerns to consider when using AI for marketing in your small business.

 

What is ChatGPT?

According to information on Wikipedia, ChatGPT is an Artificial Intelligence chatbot developed by OpenAI.  An article on Gizmodo.com noted that ChatGPT runs on “a sophisticated algorithm called a large language model.” This type of algorithm is supplied with massive amounts of data, which allows it to respond to prompts in a manner that mimics human conversation.  GPT stands for “generative pre-trained transformer,” a specific type of language model that learns as it interacts with user-initiated prompts. OpenAI released ChatGPT in Beta based on the 3.5 version in November of 2022 and  ChatGPT 4.0, the current iteration, on March 23, 2023.

 

How Can I Use Artificial Intelligence?

Since ChatGPT hit the scene, it’s become a buzzword. ChatGPT and other comparable AI chatbots (such as Rytr, Copysmith, Jasper, and many more) have had significant impact on content creation. Today, businesses use AI to create blog posts, newsletters, social media posts, outlines, title suggestions, and more. These businesses save time, money, and resources in the process.

In addition to writing blog posts and newsletters, ChatGPT has the capability to debug computer programs, write cover letters and business pitches, compose music, answer test questions, play games like tic-tac-toe, and so much more.

However, as with most innovation, AI created a host of legal landmines to navigate. And as a small business owner you may experience fear or concern over the ramifications of the use of AI in your business.

 

What are the Potential AI Legal Landmines?

1. Inaccurate, misleading, or out-of-date information.

One of the most obvious issues is that AI-generated content can be full of inaccurate, misleading, or even patently false information. Artificial Intelligence doesn’t have the ability to reason (yet), so it can’t parse out fact from fiction.  And it can’t determine whether or not a source of information in its database is a credible authority on the subject matter.

Your GPS app calculates the shortest route between home and your destination and guides you there with turn-by-turn directions.  But it is only as good as the information it is fed. We’ve had a bridge out in my town since the remnants of Hurricane Ida blew through town in September of 2021. But the GPS still directs me to turn down the road leading to that closed bridge. Apparently, nobody has fed the closure information into the GPS database. This is one of the many ways AI delivers bad information.

As of the writing of this article, ChatGPT only has information through May of 2021 available to it.  It lacks the capability to search the internet for current news the way Google does.

For example, I asked ChatGPT what the fine was for violation of the CAN-SPAM Act.  It answered $43,280 per email. Which was true in 2021.  But the current fine (as of May 2023) is up to $50,120 per email.

Inaccurate information may result in legal action against the unwary business owner who uses AI-generated content without fact-checking the results. To minimize the risk of liability, do the research necessary to ensure that the AI-generated content you use is accurate and free of errors. Because hey – you don’t want to be responsible for a robot’s mistakes, do you?

 2. Potential for loss of your IP rights.

The law does not keep up with technology. It never has.  Basically, it takes years for claims to work their way through the system to create legal precedent. So right now it is unclear who owns the intellectual property (IP) rights to content created by or with the help of Artificial Intelligence.

Questions have arisen about who owns the IP rights of original works of authorship (art, music, literature) created using AI. In some cases, the creators of AI programs claim ownership of the resulting intellectual property.   This is based on the creator’s ownership of the copyright in the code.  Users must look to the license for the AI program to understand the limitations on their rights to the output.

For example, in 2018, a group of Parisian art students created an AI-generated painting that sold at auction for over $400,000. This group, called Obvious, pocketed the proceeds with no acknowledgement of or payment to Robbie Barrat, the 19-year-old who created the bulk of the underlying code used in creating the painting. Barrat shared his code as open source, assuming users would build on it, not use it wholesale for profit. Ultimately, ownership of the intellectual property remains to be seen. The law needs time to adjust to fit a new reality.

In the meantime, however, the U.S. Copyright Office ruled that AI-created art “lacks the human authorship necessary to support a copyright claim.”  The ruling involves art work assembled by DABUS, an AI program created by Stephen Thaler.  Thaler was repeatedly denied registration by the U.S. Copyright Office and has now filed suit.  The lawsuit requests the U.S. District Court for the District of Washington D.C. to decide the novel question:

“Can someone register a copyright in a creative work made by an artificial intelligence (“AI-Generated Work”)?”

3. Possible IP Infringement Claims.

Then there’s the flip side of the coin: the possibility of being sued for intellectual property infringement.  Artificial Intelligence software relies on algorithms and machine learning to create content, but it can only do so by drawing on pre-existing works.  The pre-existing works fed into the software database for ChatGPT cover a laundry list of resources including internet articles and blog posts, books, Wikipedia, podcasts, etc. Litigation is already underway against AI developers for the use of scraped content, including text and images, in the knowledge base of AI programs.  And many legal experts believe it’s only just begun.

Do You Own the Rights to What You Create?

According to OpenAi’s own terms of use, they don’t own the IP rights to anything you create using ChatGPT. So you might reasonably assume that anything you create with ChatGPT is “yours.”  But take a closer look at the disclaimer that they provide right after saying the IP is yours:

As an AI language model, ChatGPT is designed to provide information and generate responses based on its training data and algorithms. However, any information generated as a result of your queries is considered to be your intellectual property. It is important to note that if you choose to share or publish the information generated by ChatGPT, you are responsible for ensuring that you have the necessary rights and permissions to do so.

What about the underlying content?

Which raises the question, what about the underlying content that ChatGPT uses to respond to your query? Can this AI-generated content infringe on someone else’s intellectual property rights?

I believe it can and probably does!  After all, the good folks feeding the library of knowledge to ChatGPT aren’t getting permission from all of those copyright owners to give ChatGPT the material.  How the algorithms slice and dice the material to return a result is the intellectual property (trade secrets) of OpenAI. Of course, it’s a complete mystery to me.  I can’t tell you for certain whether ChatGPT copies entire sentences or paragraphs from the material it has been fed. But I’m betting it does! This is where ChatGPT may cross the fine line between inspiration and appropriation.

My Conversation with ChatGPT

I had a “conversation” with ChatGPT to ask how it creates content and whether it checks for copyright infringement.  I appreciate the bot’s frank answer:

Me: If I ask you to write a blog post, how do you create the content of the post?

ChatGPT:  As an AI language model, I use complex algorithms and machine learning techniques to generate human-like text based on the input I receive from users. …

To create the content of the post, I draw on my vast knowledge base and language models to generate sentences and paragraphs that are grammatically correct and semantically meaningful. I can also incorporate relevant facts, statistics, and examples to support the main ideas of the post and make it more compelling and informative….

Me: Once the content is generated do you check to make sure you aren’t infringing another party’s intellectual property rights such as copyright?

ChatGPT: As an AI language model, I don’t have the ability to check for potential copyright infringement or verify the authenticity of any external sources referenced in the generated content. It’s the responsibility of the user to ensure that the content created using my language model doesn’t infringe on the intellectual property rights of others and is properly cited and referenced. If in doubt, it’s always best to seek legal advice or perform your own due diligence to avoid any potential legal issues.

Of course, it’s a tough job for a user to check on infringement when ChatGPT doesn’t provide sources, references, or citations for the information it provides.  To avoid infringing on someone else’s intellectual property rights, I advise rewriting that generated content in your own words, in your own voice, so that it sounds like you wrote it, not some robot!

4. Privacy and Data Protection.

Another legal issue to consider when using Artificial Intelligence for content creation is privacy and data protection. AI algorithms work by analyzing data and learning from it to generate content. However, this data may contain personal information that is subject to privacy laws.

It is up to companies like OpenAI to ensure that they collect and process the data used to train their AI software in a way that is compliant with applicable laws. They must also utilize appropriate security measures to protect against unauthorized access or disclosure of personal information.

But it is up to the business owner using AI generated content to ensure that what they post doesn’t contain any personal information. Better safe than sorry, right?  Because, ultimately, the business owner is legally responsible for the content they put out in the world.

5. Cybersecurity Concerns.

As with just about any good thing, in the wrong hands, it can be used for evil. Chat GPT is capable of writing very convincing phishing and malware emails. There is a significant level of concern that AI will make things easier for cybercriminals to spoof the tone and voice of legitimate companies. And that makes it harder for the average consumer to spot phishing emails.

6. Job Loss.

And of course there are concerns that AI may replace many knowledge worker jobs. If you can get AI to write your marketing content, maybe you don’t need to pay a copy writer or a social media manager to generate content for you, right?

A recent issue of the Pennsylvania Lawyer contained an article addressing concerns that AI might replace lawyers.   I wanted to test AI’s legal prowess myself.  So I asked ChatGPT to create a contract for a travel retreat, designed to protect the host/coordinator of the retreat.  Having just done this for a client, I had a fresh contract to use by way of comparison.

My expectations were very low, so I was surprised that ChatGPT returned a result that actually looked like a legal contract. But on closer review, it read suspiciously like a lot of the templates that I’ve seen sold online.  The suggested fill-in-the-blank business terms were barely adequate to set expectations between the parties. And it was missing a lot of basic legal protections.  I was pleased to learn that ChatGPT isn’t ready to replace me just yet.  It has a lot more to learn about being an effective advocate for its clients!

Best Practices For Using AI-Generated Content

Treat It As A Tool. 

Use Artificial Intelligence as the tool that it is!  AI is a fantastic idea generator when you are staring at a blank screen with no idea what to write. Or struggling to come up with a catchy title.  AI can help you get past the block.  Did you like the title for this article?  I asked ChatGPT for 10 title suggestions.  I didn’t love any of them. But I used the results as a jumping off point and built out my own version from there. The final result had little resemblance to any of the suggestions.

Use Your Own Words. 

Don’t use whatever AI returns “AS IS.”  Rewrite it in your own words and your own voice.  Keep in mind that using results from AI “as is” means you’ll end up sounding exactly like the robot that wrote it.  And with all the digital noise in the world, you’ll end up lost in a sea of sound-alikes.  Make sure it is YOUR voice your audience hears!

Fact Check.

And if you are citing anything as “fact,” do the  research and source check to be sure it is accurate before sharing.

Check for Infringement.

If you want to use something the way AI has worded it (because hey, it does return some cool results), run it through a plagiarism checker such as Grammarly before publishing. Keep in mind that the creators of these programs have left the burden of making sure there is no copyright infringement on you!

Conclusion

In conclusion, using Artificial Intelligence for content creation offers numerous benefits – it’s effective and efficient. But bottom line, AI is a tool, not the ultimate solution to your copy writing needs!  Use it as a tool, but don’t get lazy and rely on the results as is.

Remember to watch out for those legal landmines. Intellectual property, privacy and data protection, and liability are just a few of the issues to consider when using AI to write blog posts, newsletters, and social media posts. By taking these issues into account, business owners can leverage AI in a way that is both smart and legally compliant.  Don’t be AI-fraid, be AI-ware!

 

Need help navigating the legal landmines in your small business? Schedule a consultation today!

 

 

PREV

The Case for Opt-In

NEXT

The Costly Consequences of Image Theft