Frequently Asked Questions
A working reference for founders who want to understand how Serotonin Legal operates, and the legal questions they face within the AI, crypto, and fintech space.
General Questions
1.1 About Serotonin Legal
What is Serotonin Legal?
Serotonin Legal is a law firm for startups and VC funds building at the frontier of AI, crypto, and fintech. We provide legal, business, fundraising, and compliance services designed to support teams through every stage of the startup lifecycle, from formation to exit.
Every founder and firm now has access to AI tools that expand the range of what they can accomplish themselves. What they don't have is a team that came from Latham & Watkins and Goodwin Procter and experience working within and alongside public companies and investment banks. On top of that, our team advises live AI/crypto intersection deals from inside Serotonin's global portfolio while the regulations governing those deals are still being written.
This gives our team a first-hand understanding of the internal pressures founders are actually managing and ability to advise on what’s needed down the line, not just the current risks law firms are trained to see.
Who does Serotonin Legal work with?
We work with private companies and venture capital investors at every stage: pre-seed founders just getting off the ground, growth-stage companies with $1M+ in funding, and mature companies raising Series A and beyond. Family offices and individual founders navigating disputes or separations are also in scope. If you're building something ambitious that touches AI, blockchain, or the exciting intersection in between, that's where we spend most of our time.
How is Serotonin Legal different from a traditional law firm?
Three things stand out. First, our team: we came from Latham & Watkins and Goodwin Procter, so the quality of our legal work is big law-caliber. Second, speed and access: we're responsive and founder-friendly in a way that large firms structurally cannot be. Third, the cost: no billing surprises, no associate hours you didn't need, and monthly quote estimates when you want them. You get the advice without the overhead that funds it at a traditional firm.
How much does a startup lawyer cost?
Legal fees vary based on billing structure, with big law rates regularly running $700-1,200 per hour for associates and often upwards of $1,500 per hour for partners. Serotonin Legal offers hourly billing for project-based work at lower rates than large firms, as well as other payment structures that may be better suited for some startups. This includes flat fees for defined projects (incl. formations, SAFEs, and standard financings) and monthly retainers for ongoing general counsel services. We send a clear proposal and cost estimate before any engagement begins, so there are no surprises.
1.2 Services & Scope
What kind of legal work does Serotonin Legal handle?
The full spectrum of startup legal needs: entity formation, fundraising (SAFEs, notes, priced rounds), commercial agreements, employment and consulting agreements, equity compensation, cap table management, M&A, joint ventures, data privacy and regulatory compliance, terms of service, and privacy policies.
Can Serotonin Legal act as a fractional general counsel?
Yes, and for many growth-stage companies it's one of the highest-leverage legal investments they make. Our fractional GC offering means we become part of your everyday team, providing legal guidance on all matters, managing outside counsel and related fees, building relationships with your investors, customers, and vendors, and developing in-house compliance systems and policies. Think of it as an in-house GC that won't break the bank.
Does Serotonin Legal require a retainer to start working?
Not necessarily. We offer both an outsourced general counsel retainer model and a traditional billable hour model, depending on your stage and preference. Pricing is tailored to the stage of your company and what you actually need.
Common Legal Questions
DIY Tier: Legal Decisions Founders Can Handle Themselves
Do I need a lawyer for a standard contractor agreement?
Possibly not. For a straightforward consulting arrangement, a standard independent contractor agreement with an IP assignment clause, and a clear scope of work is genuinely DIY with good templates. The line shifts when the contractor is writing core product code, building anything central to your product, or operating in a jurisdiction with strong worker classification rules. At that point the template still exists, but whether it fits your situation without meaningful adaptation is the judgment call. A contractor who later claims employee status or disputes IP ownership over code that became your product is an expensive problem to unwind.
Can I form a Delaware C-corp myself?
Yes. You can form a Delaware C-corp through the Delaware Division of Corporations website or a service like Clerky or Stripe Atlas for under $1,000. The filing itself is straightforward: you submit a certificate of incorporation, appoint a registered agent, and pay the state fee. Where founders run into trouble is after formation, specifically with bylaws, initial board resolutions, IP assignment from founders to the company, and restricted stock purchase agreements. The formation is the easy part. The post-formation corporate housekeeping is where mistakes compound, especially if you plan to raise institutional capital.
Can I use a standard YC SAFE for my pre-seed round?
Yes, for a standard pre-seed angel round, the YC SAFE is designed to work out of the box. It's free, widely understood by investors, and doesn't require negotiation in most cases. The risk isn't the document itself. It's what happens when you start stacking multiple SAFEs with different caps, adding side letters, or accepting non-standard terms from a strategic investor. A single SAFE at a reasonable valuation cap for a small angel round is genuinely DIY-friendly. Once you have three or more SAFEs outstanding, you're in Challenge Zone territory and should model the dilution before signing the next one.
When does a standard NDA template stop being enough?
A mutual NDA template from a reputable source (Y Combinator, Cooley GO, or similar) works for most early-stage conversations with potential partners, vendors, or hires. It stops being enough when the relationship involves access to proprietary AI training data, model weights, or trade secrets with significant commercial value; when the counterparty is a large enterprise with its own NDA they insist on using; or when you're sharing information across jurisdictions with different enforcement standards. If someone is pushing back on your template or the stakes of a leak would materially harm your business, that's when a lawyer should review.
What is a founder vesting schedule, and can I do it myself?
A founder vesting schedule subjects your equity to a time-based schedule (often monthly over 4-years with a 1-year cliff) so that co-founders who leave early don't walk away with their full stake. You can set one up yourself using standard templates that are available online from a reputable source. For co-founder situations, vesting is highly recommended to avoid a bad leaver from retaining massive equity for minimal contributions. Investors will also expect it. The vesting terms themselves are straightforward, but pairing vesting with a properly filed 83(b) election, if applicable (see below), and making sure the restricted stock purchase agreement is correctly executed is where founders benefit from a quick legal review.
What's the difference between a SAFE and a convertible note?
A SAFE (Simple Agreement for Future Equity) is not debt. It's a contract that gives the investor the right to receive equity in a future priced round. A convertible note is debt: it has a principal amount, an interest rate, and a maturity date. SAFEs are simpler, cheaper to execute, and don't accrue interest or have a repayment deadline. Convertible notes give investors slightly more protection because of the maturity date and interest. Most pre-seed and seed rounds today use SAFEs. Convertible notes are more common when investors want the added downside protection of a debt instrument or when the company is further along.
When do I need to refresh my 409A valuation?
You need a new 409A valuation at least every 12 months, or sooner if a material event changes your company's value. Material events include closing a priced round, a significant change in revenue or business model, a key customer contract, or an acquisition offer. The 409A matters because it sets the exercise price for stock options. If you grant options below fair market value, both the company and the option holders face tax penalties under Section 409A. Most companies refresh after every financing round and annually in between. Carta and other cap table platforms also offer 409A valuations as part of their product.
Considered Risk Tier: Legal Questions That Look Simple But Aren't
Our AI vendor contract assigns all model outputs to them. Is that normal?
It's more common than it should be. Many standard AI vendor agreements (including those from major providers) include broad IP assignment or license-back clauses that give the vendor rights over model outputs, fine-tuned models, or derivative works. Whether that's acceptable depends on how central AI outputs are to your product. If you're building a product where the model outputs are the product, you need to negotiate ownership, usage rights, and training data restrictions specifically. Three important clauses to flag before signing: output ownership, model fine-tuning IP, and the vendor's right to use your data for training.
We have a SAFE with non-standard terms from a strategic investor. What should I watch out for?
Non-standard SAFE terms from a strategic investor can create problems that surface at your next priced round. Common red flags include pro-rata rights that give the strategic disproportionate follow-on allocation, information rights that may conflict with obligations to other investors, MFN clauses that let the strategic match better terms you give later investors, and side letters with board observation or approval rights. The risk isn't any single term in isolation. It's how these terms interact with your cap table and governance when a lead investor in your Series A runs diligence and asks you to clean it up under time pressure.
What is an 83(b) election, and what happens if I miss the deadline?
An 83(b) election is a filing with the IRS that lets you pay tax on restricted stock at its current (usually very low) fair market value rather than when it vests at a potentially much higher value. You must file it within 30 days of receiving the restricted stock. There is no extension and no exception. Missing the deadline means you'll owe ordinary income tax on each vesting tranche based on the fair market value at the time of vesting, which can be a significant and unexpected tax bill if your company has appreciated. File it promptly. Mail it certified.
We formed as an LLC, and our VC lead now wants a Delaware C-corp. How bad is this?
It's fixable, but the conversion has real costs and complexity. Most institutional VCs require a Delaware C-corp because LLCs create tax complications for their fund structure (pass-through income to LPs, UBTI issues for tax-exempt investors). Tax treatment depends on the specifics: a properly structured conversion can be tax-free, but it requires careful planning. The earlier you do it, the cheaper and simpler it is. If you know you'll raise institutional venture capital, forming as a Delaware C-corp from the start saves time and money.
Do I need a privacy policy before launching my app?
Yes. If your app collects any personal information (names, emails, device identifiers, location data, analytics), you need a privacy policy before launch. This is a legal requirement under California's CCPA/CPRA for users in California, GDPR for users in the EU, and various state laws that are expanding rapidly. Apple's App Store and Google Play both require a privacy policy link as a condition of listing. A basic privacy policy covering what you collect, why, how you store it, and who you share it with can be drafted from templates at the earliest stage. As you scale, add data processing agreements with vendors and update the policy as your data practices evolve.
Mandatory Counsel Tier: Issues Requiring Expert Guidance
Do I need a lawyer for a priced equity round?
Yes. Once you're negotiating a Series Seed or later with institutional investors, the documents are complex enough and the terms durable enough that experienced counsel pays for itself many times over. Liquidation preferences, pro-rata rights, protective provisions, and anti-dilution mechanics all interact with each other and with your existing cap table in ways that compound across future rounds. The negotiation dynamics also matter: a lead investor's counsel will be experienced in this, and yours should be too. This is not a category where the cost of counsel is the relevant variable.
Which M&A situations or complex transactions require outside counsel?
Almost all of them. The vast majority of acquisitions, mergers, and joint ventures involve negotiated representations and warranties, indemnification structures, earnout provisions, and closing mechanics where the drafting itself is a negotiation. One development worth flagging for AI founders specifically: RWI underwriters are now asking detailed questions about AI training data provenance during diligence. If your training data includes scraped content or material obtained without clear licensing, it can create an IP liability the underwriter won't cover and shifts risk back to you. For founders building toward an exit, clean data provenance is a diligence item to address now, not during a transaction process.
We have a token and US users. What do we actually need to do?
This depends on what the token does, how it's distributed, and whether it could classified as a security. At minimum, you need a legal analysis of whether your token triggers securities registration requirements, money transmitter licensing obligations, or AML/KYC compliance under FinCEN rules. If the token has any governance, revenue-sharing, or staking features, the analysis becomes more complex. The regulatory environment is shifting, with the SEC's evolving crypto taxonomy and pending legislation like the GENIUS Act, but "shifting" doesn't mean "relaxed." Get a formal legal assessment before launching to US users.
Can an AI agent be a party to a contract? What happens when it causes a loss?
Under current law, an AI agent cannot be a party to a contract because it lacks legal personhood. The liability falls on the humans and entities that deploy, operate, or benefit from the agent's actions. California's AB 316 has narrowed the "AI did it" defense, making it harder for deployers to disclaim responsibility for autonomous agent actions. For onchain AI products, the liability chain typically runs through the developer, the deployer, and potentially the DAO or protocol that governs the agent. The structural step founders should take now: build contractual liability allocation into your product design, not just your terms of service.
Which regulatory frameworks apply to an onchain AI product, and in what order?
There is no single framework. An onchain AI product can trigger SEC securities regulation (if the product involves tokens or investment features), FinCEN AML requirements (if it facilitates value transfer), state money transmitter laws, the EU AI Act (if it serves EU users), and an expanding set of state AI laws in the US. The order depends on your product's specific features: what triggers each framework, what your user base looks like, and where you're incorporated. The most common mistake is addressing these sequentially instead of mapping them together. A regulatory sequencing analysis early in product development prevents costly redesigns later.
What does GENIUS Act compliance actually require for a stablecoin issuer?
The GENIUS Act establishes a federal framework for payment stablecoin issuers, including reserve requirements (1:1 backing with high-quality liquid assets), registration with a federal or state regulator, regular reserve attestations, and AML/KYC compliance obligations. Issuers with more than $10 billion in circulation fall under federal regulation, while smaller issuers can operate under qualifying state regimes. The legislation also introduces consumer protection requirements and sets standards for interoperability.
If you're building or operating a stablecoin, the compliance obligations are specific and detailed enough that you need counsel to map your product against the framework's requirements, especially as pending regulations like the CLARITY Act are expected to add additional layers of regulatory complexity.
What is the "investment contract separation" doctrine, and what does it mean for my token?
The investment contract separation doctrine is the SEC's position that a digital asset itself can be distinct from the investment contract through which it was originally sold. In practice, this means a token sold in an initial offering that was a securities transaction could eventually be traded on secondary markets without each resale being a securities transaction, provided certain conditions are met. This matters for token projects because it opens a potential path from regulated offering to freely tradable asset. The conditions, timing, and required disclosures are project-specific and still evolving, so this is not a self-service analysis.
We're building an AI-assisted DAO. Who's liable when an AI-assisted governance decision causes a loss?
Liability depends on the DAO's legal structure, the role of the AI in the governance process, and whether the loss resulted from a design flaw, a failure to disclose, or an autonomous action. If the DAO is an unincorporated association, individual members and token holders may face personal liability. If it's wrapped in a legal entity (foundation, LLC), liability may be contained but still flows to the humans who designed, deployed, or approved the AI governance mechanism. The safest approach: establish a clear legal wrapper, disclose the AI's role in governance to participants, and build human override mechanisms into the governance design.
What are RWI underwriters looking for in an AI company acquisition?
Representations and warranties insurance (RWI) underwriters have started asking detailed questions about whether a company's AI models were trained on properly licensed data. If your training data includes copyrighted material, scraped content, or data obtained without proper consent, it can create an IP liability that the underwriter won't cover, which shifts the risk back to you or reduces your purchase price. For founders building toward an exit, this means documenting data provenance now: where your training data came from, what licenses apply, and whether your data pipeline has audit trails. Clean data provenance is becoming a material diligence item.
Do I need a fractional GC, or is outside counsel enough?
Outside counsel is enough if your legal needs are project-based and predictable (a financing, a specific contract, a regulatory filing). A fractional GC makes more sense when legal questions are coming up weekly across multiple areas (employment, commercial contracts, product compliance, investor relations) and you need someone who understands your business context and can make judgment calls without a briefing each time.
Curious to learn more about Serotonin Legal? —
Get in Touch
Frequently Asked Questions
A working reference for founders who want to understand how Serotonin Legal operates, and the legal questions they face within the AI, crypto, and fintech space.
General Questions
1.1 About Serotonin Legal
What is Serotonin Legal?
Serotonin Legal is a law firm for startups and VC funds building at the frontier of AI, crypto, and fintech. We provide legal, business, fundraising, and compliance services designed to support teams through every stage of the startup lifecycle, from formation to exit.
Every founder and firm now has access to AI tools that expand the range of what they can accomplish themselves. What they don't have is a team that came from Latham & Watkins and Goodwin Procter and experience working within and alongside public companies and investment banks. On top of that, our team advises live AI/crypto intersection deals from inside Serotonin's global portfolio while the regulations governing those deals are still being written.
This gives our team a first-hand understanding of the internal pressures founders are actually managing and ability to advise on what’s needed down the line, not just the current risks law firms are trained to see.
Who does Serotonin Legal work with?
We work with private companies and venture capital investors at every stage: pre-seed founders just getting off the ground, growth-stage companies with $1M+ in funding, and mature companies raising Series A and beyond. Family offices and individual founders navigating disputes or separations are also in scope. If you're building something ambitious that touches AI, blockchain, or the exciting intersection in between, that's where we spend most of our time.
How is Serotonin Legal different from a traditional law firm?
Three things stand out. First, our team: we came from Latham & Watkins and Goodwin Procter, so the quality of our legal work is big law-caliber. Second, speed and access: we're responsive and founder-friendly in a way that large firms structurally cannot be. Third, the cost: no billing surprises, no associate hours you didn't need, and monthly quote estimates when you want them. You get the advice without the overhead that funds it at a traditional firm.
How much does a startup lawyer cost?
Legal fees vary based on billing structure, with big law rates regularly running $700-1,200 per hour for associates and often upwards of $1,500 per hour for partners. Serotonin Legal offers hourly billing for project-based work at lower rates than large firms, as well as other payment structures that may be better suited for some startups. This includes flat fees for defined projects (incl. formations, SAFEs, and standard financings) and monthly retainers for ongoing general counsel services. We send a clear proposal and cost estimate before any engagement begins, so there are no surprises.
1.2 Services & Scope
What kind of legal work does Serotonin Legal handle?
The full spectrum of startup legal needs: entity formation, fundraising (SAFEs, notes, priced rounds), commercial agreements, employment and consulting agreements, equity compensation, cap table management, M&A, joint ventures, data privacy and regulatory compliance, terms of service, and privacy policies.
Can Serotonin Legal act as a fractional general counsel?
Yes, and for many growth-stage companies it's one of the highest-leverage legal investments they make. Our fractional GC offering means we become part of your everyday team, providing legal guidance on all matters, managing outside counsel and related fees, building relationships with your investors, customers, and vendors, and developing in-house compliance systems and policies. Think of it as an in-house GC that won't break the bank.
Does Serotonin Legal require a retainer to start working?
Not necessarily. We offer both an outsourced general counsel retainer model and a traditional billable hour model, depending on your stage and preference. Pricing is tailored to the stage of your company and what you actually need.
Common Legal Questions
DIY Tier: Legal Decisions Founders Can Handle Themselves
Do I need a lawyer for a standard contractor agreement?
Possibly not. For a straightforward consulting arrangement, a standard independent contractor agreement with an IP assignment clause, and a clear scope of work is genuinely DIY with good templates. The line shifts when the contractor is writing core product code, building anything central to your product, or operating in a jurisdiction with strong worker classification rules. At that point the template still exists, but whether it fits your situation without meaningful adaptation is the judgment call. A contractor who later claims employee status or disputes IP ownership over code that became your product is an expensive problem to unwind.
Can I form a Delaware C-corp myself?
Yes. You can form a Delaware C-corp through the Delaware Division of Corporations website or a service like Clerky or Stripe Atlas for under $1,000. The filing itself is straightforward: you submit a certificate of incorporation, appoint a registered agent, and pay the state fee. Where founders run into trouble is after formation, specifically with bylaws, initial board resolutions, IP assignment from founders to the company, and restricted stock purchase agreements. The formation is the easy part. The post-formation corporate housekeeping is where mistakes compound, especially if you plan to raise institutional capital.
Can I use a standard YC SAFE for my pre-seed round?
Yes, for a standard pre-seed angel round, the YC SAFE is designed to work out of the box. It's free, widely understood by investors, and doesn't require negotiation in most cases. The risk isn't the document itself. It's what happens when you start stacking multiple SAFEs with different caps, adding side letters, or accepting non-standard terms from a strategic investor. A single SAFE at a reasonable valuation cap for a small angel round is genuinely DIY-friendly. Once you have three or more SAFEs outstanding, you're in Challenge Zone territory and should model the dilution before signing the next one.
When does a standard NDA template stop being enough?
A mutual NDA template from a reputable source (Y Combinator, Cooley GO, or similar) works for most early-stage conversations with potential partners, vendors, or hires. It stops being enough when the relationship involves access to proprietary AI training data, model weights, or trade secrets with significant commercial value; when the counterparty is a large enterprise with its own NDA they insist on using; or when you're sharing information across jurisdictions with different enforcement standards. If someone is pushing back on your template or the stakes of a leak would materially harm your business, that's when a lawyer should review.
What is a founder vesting schedule, and can I do it myself?
A founder vesting schedule subjects your equity to a time-based schedule (often monthly over 4-years with a 1-year cliff) so that co-founders who leave early don't walk away with their full stake. You can set one up yourself using standard templates that are available online from a reputable source. For co-founder situations, vesting is highly recommended to avoid a bad leaver from retaining massive equity for minimal contributions. Investors will also expect it. The vesting terms themselves are straightforward, but pairing vesting with a properly filed 83(b) election, if applicable (see below), and making sure the restricted stock purchase agreement is correctly executed is where founders benefit from a quick legal review.
What's the difference between a SAFE and a convertible note?
A SAFE (Simple Agreement for Future Equity) is not debt. It's a contract that gives the investor the right to receive equity in a future priced round. A convertible note is debt: it has a principal amount, an interest rate, and a maturity date. SAFEs are simpler, cheaper to execute, and don't accrue interest or have a repayment deadline. Convertible notes give investors slightly more protection because of the maturity date and interest. Most pre-seed and seed rounds today use SAFEs. Convertible notes are more common when investors want the added downside protection of a debt instrument or when the company is further along.
When do I need to refresh my 409A valuation?
You need a new 409A valuation at least every 12 months, or sooner if a material event changes your company's value. Material events include closing a priced round, a significant change in revenue or business model, a key customer contract, or an acquisition offer. The 409A matters because it sets the exercise price for stock options. If you grant options below fair market value, both the company and the option holders face tax penalties under Section 409A. Most companies refresh after every financing round and annually in between. Carta and other cap table platforms also offer 409A valuations as part of their product.
Considered Risk Tier: Legal Questions That Look Simple But Aren't
Our AI vendor contract assigns all model outputs to them. Is that normal?
It's more common than it should be. Many standard AI vendor agreements (including those from major providers) include broad IP assignment or license-back clauses that give the vendor rights over model outputs, fine-tuned models, or derivative works. Whether that's acceptable depends on how central AI outputs are to your product. If you're building a product where the model outputs are the product, you need to negotiate ownership, usage rights, and training data restrictions specifically. Three important clauses to flag before signing: output ownership, model fine-tuning IP, and the vendor's right to use your data for training.
We have a SAFE with non-standard terms from a strategic investor. What should I watch out for?
Non-standard SAFE terms from a strategic investor can create problems that surface at your next priced round. Common red flags include pro-rata rights that give the strategic disproportionate follow-on allocation, information rights that may conflict with obligations to other investors, MFN clauses that let the strategic match better terms you give later investors, and side letters with board observation or approval rights. The risk isn't any single term in isolation. It's how these terms interact with your cap table and governance when a lead investor in your Series A runs diligence and asks you to clean it up under time pressure.
What is an 83(b) election, and what happens if I miss the deadline?
An 83(b) election is a filing with the IRS that lets you pay tax on restricted stock at its current (usually very low) fair market value rather than when it vests at a potentially much higher value. You must file it within 30 days of receiving the restricted stock. There is no extension and no exception. Missing the deadline means you'll owe ordinary income tax on each vesting tranche based on the fair market value at the time of vesting, which can be a significant and unexpected tax bill if your company has appreciated. File it promptly. Mail it certified.
We formed as an LLC, and our VC lead now wants a Delaware C-corp. How bad is this?
It's fixable, but the conversion has real costs and complexity. Most institutional VCs require a Delaware C-corp because LLCs create tax complications for their fund structure (pass-through income to LPs, UBTI issues for tax-exempt investors). Tax treatment depends on the specifics: a properly structured conversion can be tax-free, but it requires careful planning. The earlier you do it, the cheaper and simpler it is. If you know you'll raise institutional venture capital, forming as a Delaware C-corp from the start saves time and money.
Do I need a privacy policy before launching my app?
Yes. If your app collects any personal information (names, emails, device identifiers, location data, analytics), you need a privacy policy before launch. This is a legal requirement under California's CCPA/CPRA for users in California, GDPR for users in the EU, and various state laws that are expanding rapidly. Apple's App Store and Google Play both require a privacy policy link as a condition of listing. A basic privacy policy covering what you collect, why, how you store it, and who you share it with can be drafted from templates at the earliest stage. As you scale, add data processing agreements with vendors and update the policy as your data practices evolve.
Mandatory Counsel Tier: Issues Requiring Expert Guidance
Do I need a lawyer for a priced equity round?
Yes. Once you're negotiating a Series Seed or later with institutional investors, the documents are complex enough and the terms durable enough that experienced counsel pays for itself many times over. Liquidation preferences, pro-rata rights, protective provisions, and anti-dilution mechanics all interact with each other and with your existing cap table in ways that compound across future rounds. The negotiation dynamics also matter: a lead investor's counsel will be experienced in this, and yours should be too. This is not a category where the cost of counsel is the relevant variable.
Which M&A situations or complex transactions require outside counsel?
Almost all of them. The vast majority of acquisitions, mergers, and joint ventures involve negotiated representations and warranties, indemnification structures, earnout provisions, and closing mechanics where the drafting itself is a negotiation. One development worth flagging for AI founders specifically: RWI underwriters are now asking detailed questions about AI training data provenance during diligence. If your training data includes scraped content or material obtained without clear licensing, it can create an IP liability the underwriter won't cover and shifts risk back to you. For founders building toward an exit, clean data provenance is a diligence item to address now, not during a transaction process.
We have a token and US users. What do we actually need to do?
This depends on what the token does, how it's distributed, and whether it could classified as a security. At minimum, you need a legal analysis of whether your token triggers securities registration requirements, money transmitter licensing obligations, or AML/KYC compliance under FinCEN rules. If the token has any governance, revenue-sharing, or staking features, the analysis becomes more complex. The regulatory environment is shifting, with the SEC's evolving crypto taxonomy and pending legislation like the GENIUS Act, but "shifting" doesn't mean "relaxed." Get a formal legal assessment before launching to US users.
Can an AI agent be a party to a contract? What happens when it causes a loss?
Under current law, an AI agent cannot be a party to a contract because it lacks legal personhood. The liability falls on the humans and entities that deploy, operate, or benefit from the agent's actions. California's AB 316 has narrowed the "AI did it" defense, making it harder for deployers to disclaim responsibility for autonomous agent actions. For onchain AI products, the liability chain typically runs through the developer, the deployer, and potentially the DAO or protocol that governs the agent. The structural step founders should take now: build contractual liability allocation into your product design, not just your terms of service.
Which regulatory frameworks apply to an onchain AI product, and in what order?
There is no single framework. An onchain AI product can trigger SEC securities regulation (if the product involves tokens or investment features), FinCEN AML requirements (if it facilitates value transfer), state money transmitter laws, the EU AI Act (if it serves EU users), and an expanding set of state AI laws in the US. The order depends on your product's specific features: what triggers each framework, what your user base looks like, and where you're incorporated. The most common mistake is addressing these sequentially instead of mapping them together. A regulatory sequencing analysis early in product development prevents costly redesigns later.
What does GENIUS Act compliance actually require for a stablecoin issuer?
The GENIUS Act establishes a federal framework for payment stablecoin issuers, including reserve requirements (1:1 backing with high-quality liquid assets), registration with a federal or state regulator, regular reserve attestations, and AML/KYC compliance obligations. Issuers with more than $10 billion in circulation fall under federal regulation, while smaller issuers can operate under qualifying state regimes. The legislation also introduces consumer protection requirements and sets standards for interoperability.
If you're building or operating a stablecoin, the compliance obligations are specific and detailed enough that you need counsel to map your product against the framework's requirements, especially as pending regulations like the CLARITY Act are expected to add additional layers of regulatory complexity.
What is the "investment contract separation" doctrine, and what does it mean for my token?
The investment contract separation doctrine is the SEC's position that a digital asset itself can be distinct from the investment contract through which it was originally sold. In practice, this means a token sold in an initial offering that was a securities transaction could eventually be traded on secondary markets without each resale being a securities transaction, provided certain conditions are met. This matters for token projects because it opens a potential path from regulated offering to freely tradable asset. The conditions, timing, and required disclosures are project-specific and still evolving, so this is not a self-service analysis.
We're building an AI-assisted DAO. Who's liable when an AI-assisted governance decision causes a loss?
Liability depends on the DAO's legal structure, the role of the AI in the governance process, and whether the loss resulted from a design flaw, a failure to disclose, or an autonomous action. If the DAO is an unincorporated association, individual members and token holders may face personal liability. If it's wrapped in a legal entity (foundation, LLC), liability may be contained but still flows to the humans who designed, deployed, or approved the AI governance mechanism. The safest approach: establish a clear legal wrapper, disclose the AI's role in governance to participants, and build human override mechanisms into the governance design.
What are RWI underwriters looking for in an AI company acquisition?
Representations and warranties insurance (RWI) underwriters have started asking detailed questions about whether a company's AI models were trained on properly licensed data. If your training data includes copyrighted material, scraped content, or data obtained without proper consent, it can create an IP liability that the underwriter won't cover, which shifts the risk back to you or reduces your purchase price. For founders building toward an exit, this means documenting data provenance now: where your training data came from, what licenses apply, and whether your data pipeline has audit trails. Clean data provenance is becoming a material diligence item.
Do I need a fractional GC, or is outside counsel enough?
Outside counsel is enough if your legal needs are project-based and predictable (a financing, a specific contract, a regulatory filing). A fractional GC makes more sense when legal questions are coming up weekly across multiple areas (employment, commercial contracts, product compliance, investor relations) and you need someone who understands your business context and can make judgment calls without a briefing each time.
Curious to learn more about Serotonin Legal?
Get in Touch
