Whether you call it a consulting agreement, a professional services agreement, or a master services agreement, it is the document that has the most significant impact on the success of your digital transformation.
In this YouTube video, I discuss key issues to consider.
In an ERP software contract, there is a substantive difference between a warranty obligation and maintenance and support.
What those standards are and what the documentation includes is critically essential.
For every warranty obligation the ERP vendor has, you need to ensure the ERP software vendor has a responsibility to remedy that breach.
The United States Copyright Office has released a watershed report on whether and how AI-generated outputs qualify for copyright protection under current United States copyright laws. Departing from its previous policy stance, the Copyright Office now states that AI-generated outputs are eligible for copyright protection but with a key caveat: the human author requirement remains intact.
The report contains four main highlights, covered in more detail below.
AI-Generated Output Eligible for Copyright Protection
The Copyright Office concludes that, “in many circumstances these [AI-generated] outputs will be copyrightable in whole or in part—where AI is used as a tool, and where a human has been able to determine the expressive elements they contain.” The Copyright Office points to AI systems that allow users to exert control over the selection and placement of individual creative elements. As an example, it illustrates an editing process offered by Midjourney:
Although the Copyright Office will still decide on copyright protection on a case-by-case basis, people using AI tools now have a green light to seek protection for outputs if the AI tool “allow[s] users to exert control over the selection, arrangement, and content of the final output.”
Human Authorship is (Still) Essential for Copyright Protection
The Copyright Office endorsed the long-held view that copyright protection extends only to human creators, reinforcing a 2023 federal court ruling that an AI-generated image, created with no human involvement, is ineligible for copyright protection. The report reaffirms that content wholly generated by AI is not copyrightable. It also provides, however, that if a human contributes creatively to an AI-generated work—whether through selection, arrangement, or modification—those human elements may receive protection. Assessing authorship will depend on whether the human’s role involved original expression or mere instructions to an AI system, and the Copyright Office will decide these issues on a case-by-case basis.
The Key Role of Prompts in AI-Generated Content
The report details the role of text-based prompts that humans use to guide AI systems. The office acknowledged that prompts themselves, if sufficiently detailed and specific, could qualify for copyright protection, but it clarified that simple or generic prompts (such as “a cartoon spaceship”) likely do not. The report provides that even when prompts are detailed, the unpredictability of AI outputs presents challenges. The Copyright Office reasoned that because AI models operate as “black boxes” and often produce different results for the same prompt, it is difficult to establish a direct human-to-output connection akin to traditional authorship. The report concludes that while prompts themselves may sometimes be copyrightable if they demonstrate originality, the resulting AI-generated content will generally not qualify unless the user exerts substantial creative control over the expressive elements.
Current Copyright Laws Absolve any Need for Legislative Change
Some commenters advocated for new legal protections tailored specifically to AI-generated content, arguing that extending copyright to AI-assisted works could incentivize creativity. Others warned that broad copyrightability for AI outputs could undermine incentives for human authorship.
The Copyright Office ultimately concludes that existing law is well suited to address AI-related copyright issues. It maintains that AI should be treated as a tool rather than an author and that any modifications to copyright law should be carefully considered before implementation. While the Copyright Office acknowledges the need for ongoing monitoring as technology evolves, it does not recommend legislative change at this time.
Conclusion
Moving forward, the Office will continue to refine its guidance and assess the legal implications of AI advancements. But now creators and AI developers have more precise guidance on how to seek copyright protection for AI-generated output.
Recommendations
Whether AI-generated outputs infringe on copyrights is still a wide-open question.
For a cloud solution, a service level agreement is critical.
In my latest YouTube video, I discuss why these agreements are essential and the issues you must address.
In December 2024, the Illinois Supreme Court issued guidance on the use of AI (guidance) by judges and other lawyers.
The guidance comes after several high-profile instances of misuse of generative Artificial Intelligence (AI) in litigation, including cases where lawyers used AI to draft briefs that later cited non-existent case law. For instance, in several cases, lawyers used generative AI to write a brief and file it with a court, only to discover afterward that the AI “hallucinated” caselaw that does not exist. See, e.g., Gauthier v. Goodyear Tire & Rubber Co., No. 1:23-CV-281, 2024 WL 4882651, at *3 (E.D. Tex. Nov. 25, 2024) (sanctioning attorney). More responsible members of the bench and bar have used both generative and non-generative AI tools, including tools for general audiences (e.g., ChatGPT and Claude) and tools specifically designed for use by lawyers (e.g., Clearbrief, CoCounsel, and Harvey AI).
Judicial responses to AI use have varied. Some judges have proposed using AI to interpret legal terminology, while others have banned the use of AI for drafting briefs or as authority in motions. Some courts have issued standing orders requiring litigants to disclose AI usage in filings.
In contrast, the Illinois Supreme Court encourages the responsible and supervised use of AI and recommends that Illinois state court judges not require lawyers to disclose the use of AI in drafting pleadings. The guidance emphasizes that existing legal and ethical rules, rather than special provisions for AI, are sufficient to govern its use in litigation. Overall, the tone of the new guidance is to encourage the use of AI, so long as the use is responsible. The guidance also recognizes that responsibility means following the rules that already exist.
The Illinois Supreme Court’s guidance takes a relatively “pro-AI” stance. It recognizes the potential benefits of AI use, including efficiency and increased access to justice. The guidance authorizes the use of AI by litigants and judges and states that it should be expected rather than discouraged while also recommending that courts do not require litigants to disclose the use of AI in drafting pleadings.
A judicial reference sheet accompanies the new guidance, providing basic information about AI and links to other resources. The reference sheet also discusses use guidelines, like attribution and confidentiality, and identifies possible issues, such as AI hallucinations and deepfakes. (The new guidance highlights that judges “remain ultimately responsible for their decisions, irrespective of technological advancements.”)
Although “pro-AI” in the sense of not requiring specific disclosure — and discouraging an outright ban — the guidance is better understood as reiterating that the current rules governing lawyers provide sufficient regulatory structure to address the serious risks that generative AI technology poses. The guidance specifically identifies two categories of risks.
First, the guidance acknowledges concerns about authenticity, accuracy, bias, and the integrity of court filings and decisions. It makes clear that Illinois courts will be “vigilant” against AI technologies that jeopardize due process, equal protection, or access to justice. The Illinois legislature recently responded to similar concerns by restricting the use of AI in employment practices.
Second, the court’s guidance highlights privacy and confidentiality, stating that AI use must not compromise confidential communications, personal identifying information, protected health information, justice and public safety data, security-related information, or “information conflicting with judicial conduct standards or eroding public trust.”
The court’s guidance states that the Illinois Rules of Professional Conduct and Code of Judicial Conduct apply fully to the use of AI technologies. Thus, any court’s or litigant’s use of AI must comply with existing legal and ethical standards. For example, lawyers have a duty to ensure the authority they cite in court filings is accurate and good law — this includes making sure cited cases are real and not “hallucinated” by generative AI tools.
Put differently, litigators in Illinois courts still have the professional responsibility to ensure that their factual and legal citations are accurate. And they, like others doing business in Illinois, must protect the private data that their technological tools may use.
As a firm, Taft is committed to utilizing AI responsibly, and has vetted and approved specific tools, such as Clearbrief and Lexis® Create+, formerly Henchman, that improve work product while satisfying regulations and other rules. Taft’s litigators continue to monitor and develop regulations and best practices on the legal uses of AI. For questions on how to implement an AI policy or how to navigate new laws and regulations, contact an attorney in the Technology and Artificial Intelligence team.
Make no mistake—your ERP implementation or Digital Transformation will be more difficult than expected.
What are some common causes of project failure?
Keep in mind that digital transformations are not technology projects – they are business transformation projects.
I discuss these issues in my latest YouTube video.
There is an old maxim that you should never get into a contract without the ability to get out of that contract. This is even more true with ERP-related contracts.
ERP vendors like SAP, Oracle, and Infor offer discounts off the list price, expecting customers to purchase software for the duration of the contract.
Terminating an ERP software contract for breach is substantively different than terminating an ERP software contract for convenience.
Oracle is known for being aggressive during the sales cycle and dismissive of any issues or concerns arising during implementation. Oracle’s in-house legal department is often unreasonable and unwilling to compromise.
If you are considering suing Oracle for a failed digital transformation or ERP failure, make sure you know what you are getting into and have counsel who has been there before.
You must have a good case before you sue any ERP software vendor, especially Oracle. What does that mean?
While I think you should do everything you can to salvage a failing project, Oracle may be unwilling to meet you halfway. If that is the case, you may have no choice but to litigate.
Is it better to arbitrate or litigate digital transformation disputes? Like most things, it depends.
Choosing the right arbitrator(s) is everything. Pick incorrectly, and the results can be disastrous.
With litigation, disputes are governed by rules of civil procedure.
Arbitration is not always preferable to litigation and is not always less expensive.
As we often say, the Cloud is not a one-size-fits-all all proposition and it is not a silver bullet. It is a fundamental shift in how companies purchase software. It carries data security risks. If you need to customize the software to meet unique business requirements, the Cloud may not be ideal for your needs.
On-premise software traditionally consisted of a large up-front investment with a smaller ongoing maintenance obligation. With Cloud software, you are perpetually leasing software with the price increasing on an annual basis. It never goes down, and you can never stop the fee increases. Many Cloud products are not fully baked. Know what you are getting yourself into.