It’s been exciting to watch AI evolve over the past few years from a novel tool to a mainstay of modern marketing. It’s leveled the creative playing field, offered new levels of innovation, and enabled unprecedented levels of content scalability.

But as AI has become more omnipresent, it's also been met with pushback on certain legal and ethical aspects of its use. Recent lawsuits filed against major AI market brands like Meta, Anthropic, and Stability AI have raised questions about how copyright law applies to LLMs and how legal risks may translate to end users.

At Jasper, ethical technology development and transparency are core principles. These recent legal rulings have told us more about the market and legal standards for responsible AI use, and how we can allow our customers to create boldly knowing compliance is embedded into every tool and feature. Here are the key takeaways from the ruling:

1. LLM training falls under fair-use protection

When an LLM uses publicly available materials to learn patterns, it falls under fair‑use protection so long as the goal is new expression, not replication. That said, judges outlined that AI companies should be ethically sourcing all copyrighted works they use to train their models and that further compensation for original creators is on the table in the future.

2. Liability is on output

Copyright issues apply when a generated asset substantially reproduces a protected work in a way that erodes its market value (not when that work is fed to the model for training). This underscores the necessity for all AI tools to have embedded safeguards that guarantee content it generates is new and original.

3. Choosing a trustworthy provider matters more than ever

While these rulings confirmed that responsibly-trained AI models fall within fair and ethical use, they also opened new questions and considerations for the future. For this reason it's critical that brands select an AI provider well-versed in the legal aspects of AI model training and that lead on ethical and transparent practices. Your provider should offer:

• Transparent data sourcing and opt‑out paths. Full transparency about where training data comes from and let rightsholders remove their content.
• Contractual IP protection. Indemnities and warranties that shift copyright and uptime risk away from customers.
• Built‑in safety and governance. Filters, brand‑safe modes, and audit logs that block verbatim copying before launch.
• Human oversight and accountability. Dedicated experts that monitor model behavior for bias and misuse and retrain systems as needed.
• Continuous security, privacy, and compliance oversight. SOC 2 or ISO 27001, regular penetration tests, and teams that update controls as rules evolve.

Jasper is committed to being an industry leader in ethical, transparent, and innovative AI development. Learn more about how we embed these and other compliance features into the core of our offerings.

‍