Google has recently updated its policies to provide clearer guidance on how its artificial intelligence (AI) tools can be used, especially in sensitive areas like healthcare. The new rules explain that businesses can utilize Google’s generative AI to help make “automated decisions” in so-called “high-risk” fields, but there’s an important condition: a human must always be involved in some way.
To break it down: when we say “automated decisions,” we’re referring to choices made by AI based on gathered data. For instance, an AI system could decide whether to give someone a loan or assess a job applicant’s suitability. In fields where the stakes are high, such as determining health outcomes or employment opportunities, these decisions can significantly impact people’s lives.
Not too long ago, Google’s previous guidelines seemed to completely ban any kind of high-risk automated decision-making using their AI. However, Google clarified that while there are restrictions, businesses have always been able to use their AI tools for these types of decisions, as long as there is a human overseeing the process. A Google representative explained that the need for human oversight has always been a part of their policies, but they are now reorganizing the information to make it clearer for users.
In comparison, other major AI companies have stricter rules regarding their AI technologies in high-risk situations. For example, OpenAI prohibits decisions on matters like lending and hiring using their AI tools. Anthropic, another competitor, allows its AI to be involved in sensitive areas like law and healthcare, but only if a qualified professional supervises the AI’s actions, and the users need to reveal that they’re using AI.
Read Also: OpenAI States No Plans For A Sora API — At This Time
The rising use of AI for making decisions that can affect individuals has raised alarms among regulators due to the potential for bias. Studies have indicated that AI systems making credit and loan decisions can sometimes inherit previous mistakes and discriminatory practices, leading to unfair outcomes.
Organizations like Human Rights Watch have called for a ban on what they term “social scoring” systems, arguing that these could interfere with people’s access to government benefits, invade their privacy, and create unfair profiles based on biased data.
In Europe, new laws are in place designed to regulate high-risk AI systems that affect personal decisions, such as loans and job offers. Companies using these systems must register with authorities, manage risks, ensure human oversight, and report any issues that arise.
Similarly, in the U.S., Colorado has enacted a law requiring developers of high-risk AI systems to disclose important information about their tools and provide clear summaries detailing what the systems can and cannot do. New York City has also taken steps to ensure that employers cannot use automated tools for hiring decisions unless those tools have been audited for bias within the last year.
In summary, Google’s recent policy update allows for a more nuanced use of its AI tools in sensitive areas but emphasizes the necessity of human oversight to ensure fairness and accountability.
