Since its release to the public, the White House AI Bill of Rights blueprint has led people globally to reflect and question the proposition. An expansion from Biden’s vision, the blueprint—made public in October 2022—identified five principles that intends to guide the design, use and deployment of automated systems. We speak to Rob Strechay, Vice President of Product at Snowplow Analytics about the foundations, that have the potential to either enhance or constrict the US technology industry in the coming years.
How will the Blueprint for an AI Bill of Rights proposed by the White House affect AI developers?
It is important to understand that the AI Bill of Rights is a policy document or blueprint, not a binding piece of legislation. Rather, it is a set of principles to which organisations can voluntarily comply with. So today, the Bill of Rights is unlikely to change the day-to-day work of developers. It is a necessary conversation starter on the topic of AI and ethics.
In addition to the many benefits of AI and the advances the technology has made, there are also stories of trained models discriminating against groups of people, unintended biases in algorithms denying credit based on questionable correlations, or companies making recommendations to customers based on questionable conclusions. This has created a very delicate field of intimacy and privacy. As AI becomes more widely used and more available, it’s important to talk more about how to use data, algorithms and findings in an ethical way.
How will it affect enterprise adopters:
We do not expect the AI Bill of Rights to have much impact on companies using AI, as the law has no enforcement power. The most interesting aspect will be how ISVs and hyperscaler clouds adopt or build-in guardrails at the product level.
How will it affect everyday AI users?
As an educational tool, the blueprint will help inform everyday AI practitioners and remind them of the risks and potential biases at play. The biggest impact will come from the actions ISVs and hyperscalers take at the product level to eliminate bias and provide practitioners with better quality control of their models and algorithms.
Did the AI Bill of Rights go far enough, or did it go too far?
One positive aspect of the AI Bill of Rights is that the White House is shining a light on a very important issue that needs to be brought more into the public eye. Could it have gone further? Yes, it could have, but it will not have the force of law without congressional involvement. Right now, though, I am not sure the right people are in the room to have the conversations, so I hope this initiative brings the right people to the table. The European Union published a similar AI checklist in 2019, which is currently being made law through the AI Act and the AI Liability Directive. This will impact US. companies, like GDPR did, doing business in EU countries.
Are any changes necessary? If so, what should they be?
If I had a magic wand, I would make this a national conversation that the talking heads on cable news would pay more attention to. This would help more organisations and individuals alike have more discussions about AI and ethics. The power and value of AI is well documented. As technology continues to innovate and advance, it is important to ensure that it is safe, effective and that privacy is top of mind.