Download PDF
Reflecting the reality that ChatGPT was just the beginning of the AI revolution, there are now hundreds of applications and extensions that leverage generative AI to perform many tasks. However, many critics pushing for significant adoption of generative AI do so on the premise that (1) efficiency matters, (2) faster is better, and (3) more is better. There seems to be Adopting these new technologies certainly has some significant benefits that should be recognized, but adopting them too quickly exposes the adopters to a great deal of liability. This blog post explores that potential liability.
First of all, what do we mean by hundreds of applications and extensions applying generative AI? You can visit various websites that put generative AI to work, but even easier, you can download Chrome extensions that essentially link your web browser to these various programs. With extensions, they are “plugged into” your web browser, always working in the background, processing data and running services.
So what can these generative AI applications and extensions do? A better question is—what Can not do they do it for you? Get an AI product that can record and transcribe all your meetings, summarize documents and emails, summarize news, summarize Youtube videos, synthesize complex and technical texts for easy explanation. , can translate between languages, check grammar and write sentences. Create videos and images, create PowerPoints, build websites, write code, and more.
So what’s the problem? where to start Let me emphasize, the possibilities here are just unimaginable. But using emerging technologies online, free and sometimes like a guinea pig, requires some sane awareness of the risks and potential liabilities that come with using these tools. Those risks include:
- privilege. Some professionals are so privileged that they have an innate aversion to exposing customer data to others unless they follow contractual arrangements in which privileges are held and safeguards are in place. should be there. Even if he is not bound by ethical standards or concerns about privilege, especially if he has information protected by HIPAA, FERPA, GLBA, FCRA, COPPA, ECPA, VPPA, or DPPA, he may not use the data. We have a legal obligation to protect. When you download and use AI, checking the box that allows the program to release this data can have serious legal consequences.
- privacy. The Generative AI Chrome extension may require access to user data such as browsing history and personal information. The privacy risks and responsibilities out there are wide-ranging and pervasive. But until it’s proved wrong, beyond easily observable concerns, everything feeding free online technology is going into the black boxes used by AI, the larger data sets. It is safe to assume that it is, or could be, supplied to . However, more and more municipalities are giving residents the right to have their personal information deleted, even after they have consented. Simply put, machine learning works on large amounts of data. We haven’t completely cracked the code about “machine non-learning”, so removing data from a generative AI system may not be so easy. Moreover, in the context of AI and machine learning, the human element cannot be ignored. Vulnerabilities have already been identified that allow malicious parties to “talk” to her AI model to reveal other users’ email addresses and payment information. While no system is 100% secure, an early adopter of generative AI could enter content containing other people’s personal information into her AI system, leaving the AI system alive and vulnerable indefinitely. Therefore, its impact must be considered.
- Holds a lot of data. For applications of AI that record and transcribe every meeting, or create a record of activities that may not have been recorded before, AI creates an astonishing amount of data. Forget gigabytes and terabytes. This amount of data can be easily measured in petabytes. (One petabyte (PB) equals 1,000 terabytes (TB) or 1,000,000 gigabytes (GB).)This data is relevant, could be found in litigation, and definitely of Issues that future litigants need to address. As AI evolves and becomes widely adopted, lawyers and other professionals will also have AI for AI, making it easier to handle that huge amount of data. But still that data is—your Data will be stored, bundled, sold and applied in various AI black boxes.
- Unexpected and unknown applications by employees, contractors and affiliates. Do you know where your data is going, who feeds it into AI, and what happens to that data? Also, if you are an employer or supervise other people, you should be aware that they may be experimenting with these new technologies. Addressing these issues at work is a minefield, but there are concrete steps employers can take to protect their business. How you navigate the adoption of these new technologies, and how you guide your organization in working with employees who want to adopt them, raises many questions of oversight responsibility.
- Infringement of intellectual property rights. As the name suggests, generative AI generate It has content, but it’s not a vacuum. This is done by making use of existing examples and datasets, whether entered intentionally by the user or collected automatically from his website without the author’s consent. It is done. Generating AI Chrome extensions that generate content are at risk of infringing intellectual property rights. Even if users of AI tools have no intention of violating the rights of others, machine learning models that rely on other people’s content to generate new works carry the risk of copying other people’s content without their consent. . Businesses and individuals alike devote significant resources to protecting their intellectual property rights through trademarks, copyrights and patents. It also enforces those rights. Unauthorized use or duplication of protected content may result in legal disputes and liability.
- corporate compliance. Organizations must ensure compliance with applicable laws and regulations. This includes complying with consumer protection laws, advertising regulations, data protection laws and specific regulations related to the industry or sector in which the organization operates. Many of the issues raised here have not been settled in court, so at the very least, you will have legal costs to settle the outstanding issues. If you allow generative AI to extract from your data or perform tasks, this generative AI may not be aware of all federal, state, and local laws, rules, regulations, and regulations. One thing should be noted. what you do. However, the finished product will have your name on it.
- Incorrect Information and Liability. Generative AI models can produce inaccurate or misleading output. Potential liability issues may arise for the organization responsible for creating the extension if the user relies on the information generated by the extension and suffers damage or loss as a result. But anyone who feeds information into a generative AI and claims its output without guaranteeing its accuracy will be overlooked and judged. If their oversight and judgment are questioned, they may be held accountable later.
With the AI revolution on the horizon, there is no denying that adopting these new technologies will bring significant benefits. However, there are potential legal liabilities when a human sits at a computer, presses buttons, and inverts the output as a work product. Therefore, we warn our clients not to introduce this technology too quickly and without guardrails, in order to avoid exposing the adopters to a great deal of liability.
