AI on the Edge: to Regulate or not to Regulate – that’s the question!!! - Image by Gerd Altmann from Pixabay“AI is transforming all industries, but if you’re using or selling an AI product, do you need to be regulated? It’s not always straightforward.” We are asked this question, by AI designers a lot so we wanted to explore this in more detail.

In some sectors, like healthcare, legal services or financial services, regulation is a must, depending on your role within that sector. But how do you know if your AI product falls under these rules? Which regulator is applicable? And what are the risks if you’re not regulated and/or benefits?

We’ll explain the legal steps to becoming regulated, the benefits, and the downsides. We will also explore alternatives if regulation isn’t the path for you.

Stay informed, stay protected, and keep innovating.

When is Regulation Required?

The basic principle is: Ask yourself, ‘Is the underlying product or service that you are providing, regardless of how, even if AI-generated, a regulated activity or one requiring you to be regulated?’  If unsure, take advice before doing anything.

AI products can fall under various regulatory frameworks depending on the sector. But it depends on what you are doing or how you are offering these services and to whom, as to whether you need to be regulated. Here are some scenarios:

  1. Healthcare: AI used for diagnostic tools, patient care, or medical research often needs to comply with regulations. The bodies that regulate healthcare are the Medicines and Healthcare products Regulatory Agency (MHRA) and the Care Quality Commission (CQC). However, if you are supplying the APP software only and the hospital staff are carrying out the diagnosis it is unlikely you need to be regulated as the App provider.
  2. Financial Services: AI solutions for trading, financial advice, or fraud detection may need approval from the Financial Conduct Authority (FCA). The FCA has stringent rules about AI algorithms impacting financial decisions. This is a complex area as financial promotions and introductions can be regulated. If you are not advising and only offering a platform for them to themselves offer services and people locate companies on this ‘marketing platform’ you may not require regulations as you are not endorsing, promoting or advising, but there is a very fine line.
  3. Data Protection: If your AI product processes personal data, it may fall under the purview of the Information Commissioner’s Office (ICO), particularly with the General Data Protection Regulation (GDPR). Again, though, if you are only offering up the technology and not actually processing anything yourself, this might not need regulations. But if you store the data in any way, even for third parties, this opens you to data protection regulations.
  4. Legal Tech: If you are providing a template solution without any legal advice so the customers are self-populating this, without any legal or expert advice ( it’s a document-only production transaction) regulations will not be required. However, if the platform gives professional advice and guidance, amends the documents for their known commercial objectives, helps negotiate the documents so more viable, this is a regulated activity and will require SRA or Bar Council regulations or similar.

As you can see, it’s not as straightforward as you might imagine.

Legal Steps to Become Regulated

Once you’ve determined that your AI product requires regulation, the next steps typically involve:

  1. Identifying the Relevant Authority: Depending on your sector, the regulating body could be the FCA, ICO, SRA or another industry-specific authority.
  2. Preparing for Compliance: This means aligning your product with all applicable legal standards, such as data protection laws or industry-specific rules. You may need to adjust your algorithms, ensure transparency, or provide detailed reporting on your AI’s functions.
  3. Submitting an Application: You’ll need to apply for the relevant licenses and certifications. This can be a time-consuming process, often requiring documentation of your AI’s operation, risk assessments, and compliance measures.
  4. Ongoing Monitoring: Even after becoming regulated, you’ll need to keep up with evolving regulations and maintain regular reporting.

Often taking out insurance and appointing compliance officers in the company goes hand in hand with these requirements.

Risks of Not Being Regulated

Failure to comply with regulations can have serious consequences, including:

  • Fines and Legal Penalties: Regulatory bodies can impose substantial fines or even halt your product’s sale if it fails to meet legal standards.
  • Reputational Damage: Lack of compliance can hurt your brand’s reputation, as customers may lose trust in your product’s safety or reliability.
  • Loss of Market Access: Certain sectors will only work with regulated products, meaning you could miss out on significant business opportunities.

Benefits of Being Regulated

There are several advantages to pursuing regulation, even though it may seem like a hurdle:

  • Trust and Credibility: Being regulated signals to your customers that your product meets high standards of safety and ethics.
  • Access to New Markets: Many industries, especially healthcare and finance, require regulated products, so compliance opens doors.
  • Risk Mitigation: Regulation helps ensure your product is legally protected from future claims or scrutiny.

Downsides of Regulation

However, there are some drawbacks to regulation, including:

  • Cost: The application process can be costly, and staying compliant often requires ongoing investment.
  • Time: Gaining regulatory approval can be time-consuming, delaying your product’s launch.
  • Innovation Limitations: Strict regulations might stifle your ability to innovate quickly, as every change needs to be reviewed for compliance.

Alternatives to Regulation

If regulation isn’t mandatory, you can consider alternatives like:

  • Self-Regulation: Adopting internal ethical standards or guidelines can help build trust without formal regulatory oversight.
  • Voluntary Codes of Conduct: Some industries offer voluntary standards that allow companies to showcase their commitment to responsible AI without mandatory regulation.
  • Third-Party Certifications: You could opt for third-party auditing or certification to demonstrate your AI’s reliability without the full regulatory burden.

Conclusion

Whether or not your AI product needs to be regulated depends largely on its application and the industry it serves. By carefully considering the legal landscape, the benefits of trust and credibility, and the potential costs or delays, you can make an informed decision that best serves your business.

Remember, regulation isn’t just a hurdle—it’s an opportunity to solidify your brand in a growing AI-driven market.


ACLF logoA City Law Firm Limited is a leading entrepreneurial law firm in the city of London, with a dynamic and diverse team of lawyers. It was been awarded many accolades for his innovative work with emerging technologies, founders initiatives and scale-up support. They specialise in tech law, scaling and investment ready business law, IP and commercial litigation. They offer bespoke, specialist, friendly advice and support at competitive prices.

LEAVE A REPLY

Please enter your comment!
Please enter your name here