AI Can Assist, It Can’t invent

Home / Podcast / AI Can Assist, It Can’t invent

AI Can Assist, It Can’t invent

Artificial intelligence has reshaped nearly every industry. It has changed how founders run lean teams, how developers build products, and how small businesses compete with companies many times their size. But as quickly as AI tools have become essential, 2025 delivered a critical reminder from courts, lawmakers, and regulators around the world. AI can assist, but AI cannot invent.

That single rule has brought clarity, confusion, opportunity, and risk all at the same time. In a recent episode of Letters of Intent, Carbon Law Group attorneys Pankaj Raval and Sahil Chaudry broke down the state of AI and intellectual property with one clear message for small businesses. You can use AI, but you must document and prove the human work behind your innovation.

For many founders this is not just a legal detail. It is a business risk. It can impact whether your product is protectable, whether your valuation holds up during investment, and whether your idea can withstand competitors who can now copy faster than ever before.

Below is a founder-friendly guide that explains what the human-only inventor rule means, why AI compliance now affects due diligence, and how businesses can protect their most valuable asset. their ideas.

A man wearing headphones and glasses, speaking into a large professional microphone during a Riverside.fm video recording. He is one of the attorneys featured in the podcast.
A snapshot of attorney Sahil Chaudry, Associate Attorney of Carbon Law Group, during the Letters of Intent podcast episode discussing AI and intellectual property.

The Human Only Inventor Rule Explained

In 2025, courts around the world reaffirmed something that had been debated for years. Only natural persons can be listed as inventors on a patent. This became headline news when the Thaler v. Vidal decision made it clear that naming an AI system as an inventor would not be accepted by the United States Patent and Trademark Office. Europe, Australia, and several Asian jurisdictions echoed the same position. If no human invents it, it is not patentable.

For small businesses, this matters more than many realize. The moment AI models became capable of designing molecules, generating product designs, optimizing code, or creating new workflows, founders faced a new question. If AI generates the solution, who is the inventor? And can that invention be protected?

Courts have now answered. AI cannot be named. AI cannot be credited. AI cannot own or originate intellectual property. This means the human directing the process is the true inventor.

But what counts as directing the process?

Courts have suggested that it must involve more than simply giving an AI tool a prompt. Human involvement must include judgment calls, refinements, evaluations, design choices, strategic decisions, or problem solving that shapes the final output. In other words, the human must be meaningfully involved.

Imagine two scenarios.

Scenario 1. A designer tells an AI tool to create ten prototype sketches for a consumer product. The designer reviews all ten, selects one, adjusts proportions, changes materials, and redesigns the ergonomics.
Here, the designer is the inventor.

Scenario 2. A founder asks an AI tool to generate a complete mechanical design for a wearable device. The founder does not meaningfully edit the structure or make any technical decisions. They simply submit the design as is.
Here, there may be no inventor that meets the legal threshold for human involvement. The invention may be unpatentable.

This shift has turned R and D on its head. For decades, businesses operated with the assumption that if you created something novel, you could protect it. Now, the question is whether you, as a human, created it enough.

This moment is especially important for small and growing companies because patents can create a moat that protects early market share. A small startup often relies on its intellectual property to compete with larger players. Without patent protection, competitors can copy the same innovation faster, cheaper, and with less risk.

The human only rule also influences how businesses must document their work. This includes keeping detailed records of brainstorming sessions, prototypes, decision logs, and human edits made throughout the development process. Many founders overlook this. They rely on files scattered across emails, Slack channels, and different AI tools. But in the new legal era, documentation is proof. And without proof, there is no protection.

This is why businesses are shifting from casual AI use to structured AI use. They need systems that track human prompts, human decisions, and human edits every step of the way. They need invention assignment agreements that include AI specific clauses. They need internal policies that clarify what employees can and cannot use AI for.

Courts are clear. Only humans can invent. The question that now determines protection is whether a human can prove they actually did.

How AI Is Disrupting R and D And Why It Matters For Small Businesses

AI has created a strange version of the future. On one hand, businesses can now build products faster than ever before. On the other hand, so can everyone else. This speed has transformed software development into something similar to fast fashion. As Pankaj explains, Sam Altman once described the new reality as fast fashion for SaaS. It is quicker, easier, and cheaper to copy digital products than ever before.

Think of how fashion brands release entire collections every few weeks. That same cycle now exists for software. A competitor can see your product update on Monday, recreate the features using an AI code generator by Wednesday, and launch a comparable version by Friday.

This is not science fiction. It is happening every day.

Because of this rapid pace, small businesses face three new challenges.

Challenge 1. Your competitive advantage now has a shorter lifespan
If your product can be copied in days rather than months, then your market lead shrinks dramatically. Small businesses used to rely on their secret sauce. Now they must rely on legal protection, quality relationships, and customer loyalty.

Challenge 2. Your patent strategy becomes more important
Software is notoriously hard to patent. AI generated software is even more complicated. Without clear proof of human involvement, your patent application may be rejected. And once you lose the chance to file, you cannot get that moment back.

Challenge 3. Investors and acquirers are becoming more skeptical
When a company’s code can be replicated quickly, the conversation shifts from patents to defensibility. Investors ask questions like.

  • Who built this?
  • How much AI was used?
  • Is this IP actually protectable?
  • If not, is the company still worth the valuation?

A strong IP portfolio increases valuation. A weak IP portfolio decreases it. And with AI involved, founders must be ready to prove everything.

The Fast Fashion For SaaS Effect

This idea describes how fast cloning has become in the tech world. A small development team can use AI tools to reverse engineer features at incredible speed. For example, imagine a startup spends eight months building a project management tool. They ship a unique feature that helps users categorize tasks intelligently. It gains traction. It gets attention.

Within weeks, three competitors could ship nearly identical versions simply by inputting your interface and descriptions into an AI agent. These rivals may not have the original vision, but they now have your functionality.

There are two major consequences.

First, copyright may not protect you.
Copyright protects expression, not ideas. If your competitor rewrites the code, even if it performs the same task, it may not be an infringement.

Second, trade secret protection becomes harder.
AI models can infer functionality from public demos. Even if your code is secret, your idea is not.

This is why founders must adapt. The new competitive advantage is not simply what you build. It is how fast you secure protection, how clearly you document human involvement, and how well you structure your agreements. It is also how well you control access to your systems.

Small businesses that treat AI like a tool, not a replacement, are thriving. They use AI to streamline processes but keep humans in charge of invention. They maintain strong documentation. They update their employment agreements. They refine their internal policies. And they approach their intellectual property not as an afterthought, but as a core business asset.

When your competitors can clone your features, your best defense is to ensure they cannot clone your IP.

How AI Compliance Is Now Affecting Due Diligence

This is one of the most important developments of the year, yet one many business owners have not heard about. AI use is now part of due diligence. It is being baked into every serious transaction, whether it is an investment round, an acquisition, or a merger.

This means that when someone evaluates your business, they are no longer simply asking about revenue, team strength, or product roadmap. They are asking how much AI was used to build your intellectual property. They want to know whether your patents are valid. They want to know whether the copyright issues are clean. They want to know whether your training data could expose you to liability.

If your IP portfolio is not clean, your valuation may drop. And in fast moving markets, the drop may be significant.

Below are the main reasons why AI compliance has become a mandatory checklist item for deal makers.

1. Investors Need To Know Your IP Is Actually Yours

Imagine a company that uses AI to generate code for its flagship product. There is minimal human involvement. The founder cannot articulate what part of the invention the human contributed. When an investor reviews the patent applications, they ask the critical question. Who invented this?

If the answer is unclear, the patent becomes vulnerable. And when the patent becomes vulnerable, the company becomes less valuable.

This scenario is more common than founders realize. Many startups rely heavily on AI tools to speed up development. There is nothing wrong with that, but without documentation, they create uncertainty. And uncertainty is the enemy of investment.

2. Buyers Want To Avoid Hidden Copyright Liability

AI models are trained on massive datasets. If a model produces something too similar to an existing copyrighted work, it could expose your business to infringement claims. This risk increases when AI is used without guardrails.

For example, one company hired a marketing freelancer who generated blog posts using an AI system that was not trained on licensed content. Months later, a competitor accused them of using copied text. They had no idea the content came from a potentially infringing source. They simply trusted the contractor.

When the company went through a transaction, the issue surfaced during the IP audit. The buyer negotiated a reduced price due to potential liability. What could have been a small issue became a seven figure valuation impact.

This is why businesses need policies that address:

  • acceptable AI tools
  • acceptable prompts
  • acceptable use cases
  • prohibited use cases
  • required human review procedures

Without these, your team may unknowingly create legal risk.

3. Patent Offices Now Ask About AI Use

The USPTO and other patent offices have begun asking applicants whether AI was used in the invention process. They want to know whether the invention has enough human involvement to qualify for protection.

This means your application must be prepared with clarity. If you cannot explain the human role, your chances of approval weaken.

Small businesses that rely on AI must prepare early. Not after the invention. Not after the prototype. But from the moment ideation begins.

4. Contracts Must Reflect AI Ownership

Your employment contracts and contractor agreements must clarify who owns AI assisted inventions. Without this, your company could lose ownership rights if a team member creates IP using AI and the contract does not assign those rights to the business.

This issue is most common in startups that hire independent contractors. They often assume that anything created for the company belongs to the company. But unless the agreement specifically includes AI assisted work, ownership may be disputed.

This is why updating contracts is not optional. It is necessary.

The Founder Friendly Checklist To Protect AI Assisted Inventions

Pankaj and Sahil emphasize a simple but powerful checklist for founders heading into 2026. While large corporations have entire legal departments to manage risk, small businesses often overlook these steps. And in the AI era, overlooking small legal steps can create big problems later.

Here is the step by step checklist with real world examples for clarity.

Step 1. Document The Human Role In Every AI Assisted Creation

This is the foundation of your IP protection.

You need to record:

  • Who created the idea
  • Who evaluated the outputs
  • Who made the decisions
  • Who edited drafts
  • Who refined the invention
  • Why certain choices were made

If a human directed the process, you must be able to prove it.

One simple method is maintaining a development log. This can be done in Google Docs, Notion, or any internal system. Each time AI is used, team members record their involvement. This can later be used to support patent filings or investor due diligence.

Step 2. List Humans, Not AI Systems, On Patent Applications

This rule is absolute. It is not negotiable. If AI appears anywhere as an inventor, the application will be rejected.

Your patent attorney will guide you, but your internal documentation must make clear who performed the inventive steps. If that person leaves the company, you still need their assignment agreement to ensure the invention belongs to your business.

For small teams, this means you must keep updated agreements. For growing teams, this means you must have a standardized onboarding process that covers invention assignment.

Step 3. Update Your Agreements To Address AI

This includes:

employment agreements
contractor agreements
invention assignment agreements
joint development agreements
vendor agreements

These documents should specify that:

  • all AI assisted inventions belong to the company
    employees must disclose AI use in development
    prohibited types of AI use are clearly listed

This protects your ownership rights and prevents disputes.

Step 4. Train Your Team

Your team must understand:

  • How to use AI
  • When to use AI
  • When not to use AI
  • How to document AI use
  • How to review AI outputs

Training reduces risk. It also increases consistency. A small business with clear AI policies looks more professional and more attractive to investors.

Step 5. Be Transparent With Investors And Regulators

If you are asked whether AI was used, answer honestly and clearly. Investors appreciate transparency. Regulators require accuracy. Cover up creates risk. Clarity builds trust.

Transparency also speeds up deals. When a buyer reviews your IP portfolio, having clean records avoids delays.

Move Fast, But Bring Your General Counsel With You

The old startup motto was move fast and break things. Today, the motto must be move fast and break things, but bring your general counsel with you. AI empowers founders to innovate at record speed, but innovation without protection can place your entire business at risk.

Human authorship is still the key to intellectual property. AI can assist. AI can accelerate. AI can inspire. But AI cannot invent.

Small businesses that treat intellectual property as an asset will win the next decade. They will build defensible products. They will navigate due diligence with confidence. They will attract stronger investors. They will scale with stability.

This is the moment to update your contracts. This is the moment to train your team. This is the moment to document your inventions. And if you want guidance, Carbon Law Group is here to help you protect what you create.

🔗 Learn More
Website: carbonlg.com
  • Share this Podcast
  • Get in touch with us
Contact Us

First to Use vs. First to File: Global IP Explained

Succession IRL: Takeaways From This Hostile Takeover of Hollywood

The $1.50 Hot Dog Sues the White House: Why It Matters to You

The “Shadow AI” Threat in Your Business

Popular Keywords

Categories

Total Results

No Record Found

View All Results

AI Can Assist, It Can’t invent

Pankaj Raval (00:05)
Welcome back to Letters of Intent, the podcast for risk takers and deal makers. I’m Pankaj Raval founding partner of Carbon Law Group, located here in Los Angeles, but serving clients globally. Today, I am joined by our corporate attorney, Sahil Chaudry my co-host, you guys have all heard him many times before on this podcast, who leads our corporate practice here at Carbon Law Group.

Sahil (00:23)
Pankaj, it’s been a huge year for innovation and controversy when it comes to AI. AI has transformed everything from art to biotech, but 2025 also brought a hard legal truth when it comes to patents and inventorship. Courts and patent offices across the globe doubled down saying only humans can be listed as inventors. So today we’re asking, what does that mean for innovation and how should founders and creatives protect their ideas in the age of AI?

You know, this whole concept of inventorship is brand new when it comes to AI. It’s an issue that’s dominated the legal headlines all year. For people who haven’t been following very closely, what have the courts said so far?

Pankaj Raval (01:04)
been a fascinating year. feel like AI, the topic, the discussions about AI have kind of reached a crescendo.

There probably is going to be even more talk in 2026 about what this means. I think we’re still all trying to wrap our heads around it, right? There’s probably a lot more questions and answers right now, but as we see in courts are slow to get to this, but they do get to this. They do address these issues and the law evolves, and that’s what we’re seeing today. In one of the IP stories of the year around the world, courts reaffirmed that only natural people, natural persons, not AI systems can be inventors under patent law.

is a huge implication because we’re seeing how much art, how many developments that AI is producing, but still courts are maintaining that, like in Thaler versus Vidal, it’s individuals, it’s people that need to be authoring inventions for it to be protected under patent law and not the AI.

Sahil (01:53)
So basically AI can insist, but it can’t invent. A human has to guide the process, providing direction, editing outputs, making judgment calls. That’s the true inventor in the eyes of the law. okay, let’s talk impact then. For startups and R &D companies that rely on AI to innovate, how big of a deal is this?

Pankaj Raval (02:13)
It’s huge because now you have to kind go back.

and look at your whole process. You have to reevaluate the whole system, the whole process of building a patent portfolio. What is it going to take? What does that mean in terms of what is the AI involved? Many companies assumed if their AI designed a product, discovered a drug, wrote a piece of code, they could automatically protect it with patents, but the rulings have flipped this assumption. So if you can’t clearly show there was a significant amount of human involvement here, your invention may not be patentable.

And that

means you’re not gonna have IP protection, you’re not gonna be able to create a moat around that product, people are gonna be able to copy it, which is what we’re gonna be seeing. I read great tweet that my co-founder legal tech shared with me a while back, and he was saying Sam Altman came a tweet on X, I don’t know if it’s called tweet, I guess a message on X now, know what it’s called anymore, X and X and on X.

Sahil (03:00)
And X on X. Yeah.

Pankaj Raval (03:06)
Stupid Elon, why did you have to change

Sahil (03:06)
Yeah. Yeah.

Pankaj Raval (03:08)
that? But he essentially said that we’re entering the era of fast fashion for SaaS, which is so interesting because now we’re seeing all these SaaS products being copied because generally speaking code in general doesn’t have a patent protection. So now we’re seeing so much copying out there. So the question is, what are you creating? Is it How much IP has been used? gonna be a lot of questions

a lot of probably office actions issued by the USPTO to understand what actually was human derived and what was AI.

Sahil (03:37)
So this just adds such an interesting layer when it comes to diligence. mean, we often tell our clients how valuable protecting your IP is because when it comes, you know,

You want to avoid any kinds of infringement claims. And also when it comes time to raise private equity or IPO, when you’re raising capital, this is a core asset that you’re selling. So not getting this right can, I feel I can poke a lot of holes in any kind of potential sale or exit. So let’s get practical because a lot of our clients are going to be dealing with this problem. If I’m a founder using AI to develop products, what do I actually need to do to stay protected?

Pankaj Raval (04:12)
There’s actually three big things you’ve to be thinking about as a founder.

as a business owner, trying to create an IP portfolio, trying to create a moat around your That is, first of all, document the human role. Be really clear about what the human did, the author did in the development of the product or IP software, whatever it might be, because that’s gonna be really important to proof of human authorship.

list the humans on the patent. You can’t be listing AI. Hopefully people aren’t saying, also, ChatGBT created this, because that’s not going to get you a patent. You need to make sure that there’s a human listed on the patent and that human had a impact and of creating that IP. And lastly, update your contracts. Make sure your contractor agreements say that AI-assisted inventions belong to the company.

So if, you know, any contractors or employees are using AI to develop products you want to make sure that you know anything they use even if they’re developing using AI, that is still part of the company so that means your invention assignment agreements may need to be updated means other contracts other Contractor agreements need to be updated because you want to make sure that anything they created even using AI is still owned by the company

Sahil (05:20)
So it sounds like we’re saying, courts might not uphold the non-human elements of some kind of creation, but in your agreements, you wanna have that blanket umbrella style protection that’s wrapping in all of the IP so that you can get as close to the line as possible of total ownership. ⁓ So I think that seems like a really critical point, and I hope.

Pankaj Raval (05:36)
Right.

Exactly.

Sahil (05:45)
our listeners are taking note that as you’re developing your IP or you’re working with third parties, any kind of AI assisted inventions need to be wrapped in. you know, I’m wondering now, we also saw a very similar theme in copyright this year. talked about patents up until now, but with copyrights, you know, let’s say we’re talking about a movie script.

no human author, then there’s no protection. So the US Copyright Office has rejected AI created works that lacked any kind of meaningful human input. How far do you see that impact it comes to copyright?

And from the deal side, this creates a new layer of due diligence, right?

Pankaj Raval (06:23)
Yes, exactly. So, you know, when we’re talking about deals, you know, because we law firm for deal makers and risk this is letters of intent. So, you know, we’d be remiss not to mention how this impacts a lot of the deals going buyers or investors are company that has a significant IP portfolio, they need to ask these questions about, okay, what is the amount of AI used to product? Because, it’s fully AI developed,

one person who just used AI to develop the whole thing, it could be susceptible to being copied very easily and they won’t be able to prevent that. So they need to think about of AI being used to develop this IP and is it something that they can really protect? Because if not, they need to factor that into their valuation or into their approach this company.

Sahil (07:06)
So add that to our due diligence list. AI compliance is now a critical part of due diligence for any kind of transaction. ⁓ And this applies to patents. This also applies to copyright. We saw the same rule. No human author, no protection. The US Copyright Office has rejected AI-created works that lacked meaningful human input.

Pankaj Raval (07:15)
Absolutely.

Absolutely, absolutely. Yeah, I reinforces the same principle in copyright as well. You know, we see so many images nowadays being created by You know by Grok, by ChatGPT. It’s pretty amazing, you know, I mean I use it all the time too. I created my daughter’s invite, you know for her birthday party with it, you know, we there’s so many great use cases, but there’s still cases to be resolved about you know,

it be infringing on other people’s copyrights? AI was trained on other people’s work. So to the extent that other people’s work shows up in the works that you’re creating, if it is for commercial use, you’ve got to be aware of that. you’ve got to be careful. think we touched that on that in prior using services like Firefly with Adobe or Envato, ones that actually have licensed work that they use to train the models. Because if not,

could be a risk for you also with your company. We still haven’t seen how those, all those cases are going to play out, but there’s definitely a risk there. We actually helped someone in the early days of, ChatGPT last year or a year before resolve a case where side alleged that they were copying, they essentially used their blogs and repurposed their blogs and my client repurposed their blogs. And it was all because, you know, AI had,

used some of the same language for that blog on my client’s blog, but they didn’t know, are risks you gotta take. I mean, this is a risk you do take when you’re AI, but you’ve gotta be aware of those risks. So this exact same issue’s gonna be coming up in copyright, and you gotta understand, okay, what’s the level of human creativity here, and how much does that matter to the value of the work you’re putting out?

Sahil (08:50)
So let’s close with some practical advice for our listeners, founders, investors, creators who are navigating AI right now. What should they be doing? What’s the checklist?

Pankaj Raval (08:59)
So here’s the checklist for 2026. You gotta be aware that the world is changing fast and you gotta be ready for it. You wanna be on the offense and not on the defense because I’ll tell you the defense is helping a lot of clients because they didn’t take proper precautionary steps. That’s why also we have a subscription agreement for a lot of clients that they’re part of now, where they come to us ahead of time before these issues arise. It’s our general counsel service agreement on our website. We can definitely share more about that.

to be making sure that they are looking at what is the level of human input in this intellectual property product or service. It’s got to be human involvement. It’s got to be human first developing it. Next is update your agreements. Make sure your agreements address the use of AI in the development of IP. Next is train your teams. Make sure your whole team is trained on the use of AI. I go over this with my team all the time. Not only just better prompting and how to prompt, but also

to record inventorship, to record authorship to make sure it’s clear the human was involved the development of this IP. And lastly, is be transparent. Investors and will be asking about AI. So be ready with clear factual answers. Make sure it’s all recorded, make sure it’s all documented. that’s the beauty about today is that there’s a lot of easy, because of AI, the funny thing is there’s a lot of easy ways to document things nowadays as well.

But make sure you’re doing that and you’re taking those extra steps because it could come up in the future where, especially in the sale of your company, when there is audit, there’s gonna be questions about,

how much AI was used in the development of that IP, and that really matters in determining the value of your product or service.

Sahil (10:27)
And that determines the value of your product and service. And also when it comes time for the sale of your enterprise, add from the business side, treat your IP records like an asset because they are. We insist on this with all of our clients. You need to know that your trademarks, your copyrights, your patents, your IP are extremely valuable. It doesn’t feel like that. People like to move very quickly at the initial stages of their business.

And it doesn’t feel that valuable, but trust when you get to that point where you’re ready for your exit, you are going to want to have your IP records locked in. because when you go to raise or sell those details or make your innovation. ownable and valuable. and so there you, yeah, there you have it. 2025 was the year the courts told us AI can assist, but it can’t invent. Human authorship is still the tether to intellectual property. So

Pankaj Raval (11:04)
Absolutely.

Sahil (11:16)
For 2026, remember to protect your ideas, document your process, and keep humans in the loop.

Pankaj Raval (11:22)
Exactly, Sahil. Things are changing fast. AI, we thought the world was moving faster before. AI is only speeding fast we are moving.

There’s this in the startup world that Mark Zuckerberg made popular, called move fast and break things. But you gotta be careful because if you are moving too fast and you break too much, you could be stepping on a landmine that could blow up your whole business.

Sahil (11:42)
So move fast and break things, but bring your general counsel with you. thank you all for joining us for this episode of Letters of Intent. We are here for you as you take your risks and make your deals. I’m Sahil Chaudry corporate attorney here at Carbon Law Group.

Pankaj Raval (11:47)
Yes, exactly.

And I’m Pankaj Raval, founder of Carbon Law Group and IP and corporate attorney as well. And we are here to help all businesses, whether you’re just starting out or trying to scale up, we’re here to help you from your mergers and acquisitions to to trying to figure out how to company to the next level. We’re here to guide you along the way and we’d love to be a partner in that journey.

Thank you very much. Thanks, Sahil.

Get in touch with us

Lead Form Main

The main Lead Form

This field is for validation purposes and should be left unchanged.
Name(Required)
SMS Policy(Required)