Is Web Scraping Legal? Unraveling the Complexities of Data Privacy & Ownership

How Is AI Collecting My Data?

The chances are pretty good that something you have entered online has been ingested by an AI program. This could be something as simple as a Google review, a Facebook post, or a blog or article you wrote in the past. There are many sources of data for AI tools, which include the following:

• Structured data: databases or spreadsheets.
• Unstructured data: emails, social media posts, photos, and voice recordings.
• Streaming data: data generated from sources like IoT devices, stock price feeds, and social media streams.

AI tools collect data directly from online forms, surveys, and tracking codes, and indirectly from users’ likes, shares, and comments. The software will then transform raw data by cleaning, processing, and analyzing it. The all-important data cleaning process is meant to remove or correct errors to ensure the data is accurate, consistent, and reliable.

Consumer privacy protections should be considered during this entire process, though our attorneys suspect these considerations are often forgotten in the race to create the best AI product.

The techniques and data harvesting methods used to gather this data include the following:

• Web Scraping: The most well-known way Big Tech has trained its AI is by gathering vast amounts of data by automatically pillaging content from websites. While some of this data is public, personal details may be collected without user consent.
• Biometric Data: AI systems use facial recognition, fingerprinting, and other biometric technologies gathered from smartphones, work security systems, and IoT home appliances, raising substantial privacy concerns.
• Social Media: AI algorithms analyze social media activity, capturing demographic information, consumer preferences, and other data that accumulates on an hourly basis.

AI’s Insatiable Quest for Data 

Generative artificial intelligence systems like Bard, Copilot, and OpenAI rely on all available data for optimal machine learning. These systems gobble up news stories, fictional works, forum posts, Wikipedia articles, photos, podcasts, and YouTube videos—essentially any available text, images, audio files, and videos. 

You would think companies would want to be more selective, but the principle remains: the more information you feed OpenAI, the faster it learns and the more intelligent it becomes. More than ever, these models require high-quality proprietary data. 

While the internet offers vast scraping opportunities, its resources are finite. Even with large websites like Reddit and Wikipedia growing their data load daily, companies and AI programs are using the data faster than it is being produced.

Therefore, companies that train these AI tools are always looking for more data they can harvest and feed a system. To establish themselves as leaders in AI and compete with the best products, big tech companies are on a constant hunt for digital data needed to advance the technology.

Generative AI Legal Issues

So, where is this new data coming from? Over the years, big tech companies have tried to adjust their privacy policies to use all the data they collect from users legally. However, just because information is available online does not automatically grant AI systems the right to use it. 

For example, Google has admitted to training AI models on some YouTube content. Meta has said it has integrated AI into its services to use the billions of images and videos from Instagram and Facebook to help train its models. 

Below are a few ways in which generative AI data scraping can violate your privacy and legal rights: 

• Copyright infringement: Online content like news articles, artwork, songs, and more is protected under copyright law. Using an AI system to scrape and use this content without consent from the owner of the copyright could lead to a legal claim.

• Breaching terms of service: Some websites often ban data scraping in their terms of service, so if an AI program scrapes this website, it could be directly violating this contract. 

• Violation of privacy rights: Many state and federal laws require notification, consent, and the ability to opt-out regarding people’s personal information. A company could be liable if an AI model fails to follow these laws or regulations.

• Computer Fraud and Abuse Act (CFAA): This law oversees unauthorized access to computer systems and data. If an AI system scrapes data from a website with authentication or a paywall, it could violate the CFAA.  

The Lyon Firm’s privacy lawyers are investigating several complaints from individuals and companies who claim the web scraping tactics of companies have been intrusive and violated their privacy rights. Some questionable data-gathering practices have prompted lawsuits over copyright and licensing.

Companies With a Web Scraping Lawsuit Filed Against Them

Several companies are caught in the middle of the debate to answer: is web scraping legal? Such examples where companies’ use of AI to harvest data has landed them in hot water include: 

• A class action AI privacy lawsuit was filed against GitHub, Microsoft, and OpenAI regarding Copilot, a tool that predictively generates code based on what a programmer has already written. Plaintiffs allege Copilot used code and failed to provide attribution to the original programmer.

• A class action AI lawsuit was filed against Stability AI, Midjourney, and DeviantArt, with plaintiffs alleging the programs directly infringe on individual copyrights by training the AI on works created by the plaintiffs.

• Getty Images filed another complaint against Stability AI for allegedly copying and processing millions of images owned by Getty in the U.K.

• Many authors are suing OpenAI for allegedly infringing on authors’ copyrights. The complaint estimates that hundreds of thousands of books were copied into OpenAI’s training data. In a similar case, The New York Times is suing OpenAI for copyright infringement, with allegations that millions of articles were used to train and develop OpenAI’s chatbot.

• A class action AI lawsuit was filed against Google for allegedly misusing personal information and copyright infringement. The lawsuit says photos from dating websites, Spotify playlists, TikTok videos, and books were taken without consent to train Bard.

Examples of AI Data Lawsuit Settlements

Multiple companies have been ordered to pay large sums for allegedly illegal data scraping, including: 

• In June 2024, Clearview AI agreed to pay $50 million to compensate plaintiffs who accused the company of scraping publicly available images on websites like Facebook, Venmo, and millions of others.

• Vimeo agreed to pay $2.25 million to certain users of its AI-based video creation and editing platform Magisto to resolve claims it collected and stored their biometric data in April 2023. 

• Google reached a $1.6 billion settlement with Singular Computing to resolve a lawsuit accusing the search giant of patent infringement during its development of artificial intelligence (AI) programs.

Copyright Dispute and AI Scraping 

The New York Times and several other newspapers have sued OpenAI and Microsoft for using copyrighted news articles without permission to train chatbots. However, OpenAI and Microsoft argue that using the articles was legal under copyright law because they altered the works for a completely different purpose. 

According to a report by The New York Times, tech companies like OpenAI, Google, and Meta have allegedly ignored corporate policies, intellectual property rights, and widespread consumer privacy concerns to obtain the all-important data.

The “fair use” doctrine allows third parties limited use of copyrighted works without permission or payment. Yet, news outlets argue these companies cannot freely use their material without consent or compensation. Furthermore, in some cases, chatbots are incorrectly citing articles and falsely attributing reporting to certain newspapers.

Are There AI Lawsuits for Copyright Infringement?

Just as in the past, with new technology, innovations in artificial intelligence are raising questions about how copyright law principles will apply to content created or used by AI. Many individuals and companies have filed lawsuits over what they believe are illegal data collection practices. 

The top chatbot systems have reportedly been trained on reams of digital text spanning trillions of words. However, the most important data for any AI model is high-quality information in books and articles, which are almost all protected by copyright. 

In the race to find large amounts of data to dump into their AI model, Meta, the parent company of Facebook and Instagram, recently discussed buying the publishing house Simon & Schuster to obtain the rights to important works.

Unfortunately, these companies are not always willing to pay. Generative AI systems have copied huge amounts of data that should require a license, infringing on copyrights. The Digital Millennium Copyright Act (DMCA) provides restrictions on the removal or alteration of copyright management information.

The U.S. Copyright Office has begun to confront the issue as they have been bombarded with complaints and questions about how generative AI outputs can be copyrighted and how generative AI might infringe on existing copyrights.

Can AI-Generated Works be Copyrighted?

The question of whether copyright protection can be afforded to AI outputs hinges partly on the concept of “authorship.” The Copyright Act affords copyright protection to “original works of authorship.” The law does not clearly define who or what may be an “author.” However, the U.S. Copyright Office recognizes copyright only in works “created by a human being.”

This makes sense, of course, but when humans author a work with the help of an AI system, there are suddenly gray areas. As it stands right now, works created by humans using generative AI software can be entitled to copyright protection, but each case depends on the level of human involvement in the creative process. If AI programs are used, the individual also must disclose the use of AI, or they could violate current copyright law.

In March 2023, the Copyright Office released guidance on the use of AI and how AI-generated material may be copyrighted when “sufficiently creative” human arrangements are combined with AI-generated material.

Unfortunately, this is all rather vague, leading the courts to interpret these situations differently. Right now, there are very few past court judgments to cite, and new cases will be heard on an individual basis for quite some time.

Strong proponents of artificial technology say AI-generated works should receive copyright protection, arguing that AI programs are very similar to other tools that human beings have used to create copyrighted works in the past.

But if you assume that some AI-created works may be eligible for copyright protection, who owns that copyright? The original forces behind the coding and training of an AI program could give an AI creator a strong claim to some form of authorship. However, at the moment, Open-AI’s terms of use designate any copyright to the end user.

Who Can File a Web Scraping Lawsuit?

Recent plaintiffs in AI data lawsuits have included authors, artists, and major media organizations claiming AI programs have stolen their work without consent or compensation. Many have filed claims against companies developing programs, claiming the organizations breached copyright law.

Also, lawsuits may not only apply to outright using someone’s work. There are also potential privacy violations when generating work in the style of someone else. This protection is known as the right of publicity, which prohibits using someone’s likeness, name, image, voice, or signature for commercial gain.

Can I Join a Class Action AI Data Lawsuit?

It’s not only large media conglomerates that are concerned about the use of their original material. Thousands of trade groups, artists, and authors have raised awareness of the need for more direct, clear protections of their works regarding artificial intelligence software and tools. 

While Big Tech dolls up their data harvesting methods as “transformational” and “for the greater good,” individuals whose works have been used are calling it simply “institutional theft.”

You do not need to be a massive corporation in order to take legal action against AI companies wrongly appropriating content. There have been multiple instances where individuals have banded together in a class action to retake control of their personal information or intellectual property. 

Can I File an AI Lawsuit?

If you believe AI data collection practices have compromised your private information, you may qualify for a case. This may be the only way to protect your personal privacy and intellectual property.

Without your resolve to file an AI lawsuit, large corporate defendants will continue to invade, utilize, and even publicize the information of many without any risk of monetary penalty. 

By holding companies accountable for any blatant invasion of privacy, every individual will have more control over how their data is used in the future.