A 12 months after main American Massive Tech firms agreed to the Biden administration’s voluntary commitments to handle the dangers posed by synthetic intelligence (AI), Amazon stated there’s a want for world alignment on accountable AI measures.
“It’s now very clear we will have guidelines that shield in opposition to dangers, whereas additionally making certain we don’t hinder innovation,” David Zapolsky, senior vice chairman, world public coverage and common counsel at Amazon, stated in a Monday (July 22) article posted on the corporate’s web site. “However we nonetheless must safe world alignment on accountable AI measures to guard U.S. financial prosperity and safety.”
One method to accomplish this method to accountable AI is for all firms within the AI discipline to decide to a accountable deployment of the expertise, Zapolsky stated. This might embody measures like one Amazon has taken: embedding invisible watermarks into its device that permits customers to generate photographs, with a purpose to cut back the unfold of disinformation.
As well as, all firms must be clear about how they’re creating and deploying AI, Zapolsky stated. Within the case of Amazon, the agency has created AI Service Playing cards that inform Amazon Web Services (AWS) prospects the constraints of its AI companies, about accountable AI greatest practices and about how they’ll construct AI purposes safely.
One other key to accountable AI is collaboration and data sharing amongst firms and governments, Zapolsky stated. Right here, Zapolsky pointed to the U.S. Synthetic Intelligence Security Institute Consortium, which was established by the Nationwide Institute of Requirements and Expertise to advance analysis and measurement for AI security.
“For the U.S. and our allies to unlock the advantages of AI whereas minimizing its dangers, we should proceed to work collectively to ascertain AI guardrails which might be per democratic values, safe financial prosperity and safety, guarantee world interoperability, promote competitors, and improve secure and accountable innovation,” Zapolsky stated.
On July 21, 2023, the White Home stated that Amazon, Anthropic, Google, Inflection, Meta, Microsoft and OpenAI made voluntary commitments to assist transfer towards secure, safe and clear improvement of AI technology.
These commitments embody a variety of measures designed to raised perceive the dangers and moral implications of AI whereas offering better transparency and limiting the potential for misuse.
