Page 427 - AI for Good Innovate for Impact

AI for Good Innovate for Impact

Oeldorf-Hirsch, 2016). As a result, many users blindly agree to these terms without reading
or comprehending them (Milne and Culnan, 2004), potentially compromising their privacy or
unknowingly agreeing to unfavorable clauses. This creates an imbalance of power between
users and service providers, allowing companies to draft ToS agreements that heavily favor 4.4-Productivity
their interests while limiting user rights and imposing strict liabilities on them (Marotta-Wurgler,
2010). With the increasing use of digital services, this problem affects a vast number of users
worldwide.

Artificial intelligence (AI) offers a solution at combating the issues described above. Due to its
automation and speed, AI has the ability to process and analyze large volumes of documents
faster and more accurately than humans can. Also, AI systems have the ability to improve over
time by learning from user interactions and past performance. These features make AI a useful
tool in helping everyday users to understand and analyze terms of service in ways that have
previously proved difficult. Additionally, Machine Learning (ML) can be used to help the model
recognize patterns in legal documents, identifying problematic clauses or inconsistencies
across multiple contracts (Surden, 2014; Katz et al., 2014). Over time, it can become more
accurate at spotting unfair terms or data privacy concerns.

While AI tools for legal documents, such as LawGeex, exist, there remains a lack of AI-driven
tools specifically focused on summarizing terms of service for everyday users. These tools,
while powerful, are often tailored for businesses or legal professionals, leaving a gap in the
market for tools that cater to individual consumers. Current literature (Goodman and Flaxman,
2017) underscores that while AI is improving the legal domain, consumer protection remains
underdeveloped in AI research. The need for AI tools that simplify legal documents for the
general public, and everyday users, who often lack the time or expertise to analyze contracts,
remains largely unmet. AI systems that can offer real-time analysis of terms of service, flagging
risky clauses or comparing terms across services, could play a critical role in empowering users.

The proposed use case aims at utilizing these properties and technologies of AI in order to
create an application with an AI model that can summarize, rate and flag potentially unfair
clauses contained in the ToS agreements of various digital services in order to provide better
understanding and transparency, empowering users to make more informed decisions and
better protect their digital rights and privacy. The AI model will be a pre trained legal LLM fine
tuned on a comprehensive ToS dataset to perform the following major functions:

• summarize ToS and privacy policy documents
• rate ToS and privacy policy documents based on how well user rights are protected
• identify and flag potentially problematic or unfair clauses
Use Case Status: The use case is part of a final year research project

2�2 Benefits of the Use Case

• The proposed use case can address key gaps in fields such as consumer protection, ToS
transparency, and the application of AI in legal domains, which could help to further
advance the research in and development of these fields.
• The proposed use case encourages ethical consumption by educating users about their
right and implications of ToS and privacy policy agreements as well as encouraging them
to make informed choices about the digital services they use. It can also foster responsible
business practices by encouraging producers of these digital services to review their
terms of service and how they handle user data.

391