The report released Wednesday said the company's disclosures about its AI programs go beyond what it and other major AI companies agreed to with the White House last summer.
“This report allows us to share our maturing practices, reflect on our learnings, set goals, hold ourselves accountable and earn the public's trust,” the company said in a statement Wednesday. Stated.
“We are committed to sharing our learnings early and often and having robust conversations around responsible AI practices,” the company said.
Microsoft released the report amid continuing concerns about chatbots that rely on generative AI and sometimes return strange, alarming, and unreliable responses.
In February 2023, the company announced that shortly after releasing its Copilot chatbot, long conversations with more than 15 questions caused the chatbot to become “repetitive, not necessarily informative, or not consistent with the tone we designed.” “They may be prompted or provoked to respond in a way that they have not responded to.” ”
The company said it relies on real feedback from users to improve features.
In March, Microsoft engineers filed concerns with the Federal Trade Commission (FTC) about potentially offensive and deceptive images created using Copilot Designer, the company's AI image tool powered by OpenAI models. Did. CNBC report.
In January, the FTC began investigating financial relationships involving billions of dollars between investors and major AI companies, including Microsoft, Alphabet (Google), Amazon, Open AI, and Anthropic PBC, the FTC said. Announced.
Britain's Competition and Markets Authority is also investigating Microsoft's relationships with Mistral AI, Inflexion AI and other startups, as well as Amazon.
Congress has held multiple hearings on the possibility of putting guardrails on AI because of its potential to cause harm and disruption.
