AI in software development: IP challenges for investors

4 min readJan 15, 2024

The modern wave of AI is reshaping various industries and software development is at the forefront. Innovations like GitHub Copilot and cutting-edge language models are transforming how we code and perceive software creation. But with all transformative innovations, challenges arise. Drawing from our daily work with startups, we see that the biggest challenge today is around intellectual property rights (IP). Striking a harmonious balance between legal nuances and product development is a tightrope walk companies need to master.

Challenges in the digital era

In our digital age, we have a huge amount of data at our fingertips. But there’s a catch: not all data is free and unlimited. Some data, including source code, is protected by licenses that set limits on its use. When AI systems dive deep into these large datasets for training purposes, they may accidentally end up learning materials containing protected source code, for example, licensed under the General Public License (GPL). The AI’s interaction with such code usually doesn’t pose a legal problem on its own, the situation becomes more complicated if the AI system begins to incorporate segments of the licensed code, especially GPL-licensed segments, into its output. This is important to keep in mind when using various tools such as GitHub Copilot to help you with your code.

Why is GPL important?

The GNU General Public License (GNU GPL or simply GPL) is an important element in the world of open-source software licensing. It was authored by Richard Stallman and guarantees the public autonomy to use the program in several ways. It allows individuals to use, modify and even redistribute software without having to pay a fee. This very essence, while liberating, can also pose a significant risk. If you include a piece of GPL-licensed code in your source code, it creates a risk to investors, founders and developers. Overnight, the value of the software can drop if it’s freely distributed under the GPL, bypassing channels that generate revenue for companies. What’s the reason?

Real-world implications

The GPL domain has its specificities, one of which is the copyleft provision. IGPL is not just an ordinary license. Its reach can be extensive and even a small piece of GPL-protected code can affect your entire project. The copyleft licenses, often referred to as “viral”, means that using GPL code can subject your entire application to GPL restrictions.

One of the notorious examples is the WRT54G router from Linksys. After Linksys was acquired by Cisco, it was found that the router unintentionally incorporated GPL-licensed code, an oversight made by a subcontractor. This oversight prompted the Free Software Foundation (FSF) to urge Cisco to follow the GPL conditions. Cisco’s market for premium routers suffered due to this issue, making them redesign their firmware without utilizing GPL-licensed software to sustain their initial business model.

The IP’s role in valuation

For investors, unaddressed IP weaknesses can drastically reduce the potential of a promising venture. Even small IP missteps can cause legal troubles and damage public trust, harming profitability. For tech start-ups, especially those using AI, unexpected IP issues can lead to legal disputes, damage brand image and undermine investor trust. So, what steps should you follow to stay within the law?

Eldison’s top tips

  1. Tracking the origins of code: In our AI-driven world, tracing the origin of every piece of code is a primary concern. Eldison recommends traceability in AI-based tools by storing historical versions and having an AI policy in place at your company to help ensure preparedness throughout the investment journey.
  2. Consistent IP monitoring: Much like routine financial assessments, regular IP checks are pivotal. They serve as the first line of defense against unexpected legal battles.
  3. Divide liability with contractors: Clearly define and split responsibilities and liabilities related to AI use with any external contractors or third-party developers.
  4. Compensation agreements: Secure indemnification agreements with AI service providers to protect yourself against legal issues, especially related to intellectual property violations.
  5. Adhere to AI software Terms and Conditions: Strictly follow the terms and conditions of AI software usage to avoid legal issues and ensure ethical and compliant use.

The horizon is full of opportunities and challenges. If you are on your way to building a new software and need a legal eye to look over your IP, send us a DM. We’ll help you be compliant and investable but won’t restrict your business.