Microsoft adds OpenAI’s GPT-3 text-writing AI to its cloud computing
Microsoft announced yesterday that it will be offering an updated version of the AI Natural Language Program (NLP) GPT-3 to enterprise customers as part of its Azure cloud platform.
Why it matters: The move probably puts the most powerful AI writing and reading algorithm at the fingertips of large companies who can use it to automatically analyze and generate new written content.
Send the news: While OpenAI — the artificial general intelligence research firm that created GPT-3 — has access to the model and will continue to sell it through its own API, Microsoft will offer a version for enterprise customers that emphasizes “safety and security,” it says. Eric Boyd, corporate vice president of Azure AI at Microsoft.
- Looking back: In 2019, Microsoft invested $1 billion in OpenAI, a partnership that made it the exclusive provider of cloud computing services for the company.
How it works: GPT-3 is a natural language transformer program trained on half a trillion words on the internet, making it the largest model in the world when it was released last summer.
- Boyd says Azure customers can use GPT-3 to summarize massive amounts of customer feedback or analyze transcripts of live sports broadcasts to generate running commentary.
- While companies are already using computer vision and other AI tools, “the number of natural language use cases” is increasing [for businesses] does everything else in the shade,” he adds.
The catch: Like all NLP models, GPT-3 can incorporate bias into its training set, producing text that can be flagged with sexism, Islamophobia, and other highly human ills that can expose business users to legal and reputational risks.
- OpenAI found GPT-3 — which could also be used to generate disinformation — too dangerous to be released publicly without access restrictions.
- “to bet” [GPT-3] to production, they need things like privacy protections and built-in responsible AI controls,” Boyd says. “And those are the promises that Azure really brings.”
- Microsoft will try to avoid some of those issues with protections, including vetting customer use cases and providing filtering and monitoring tools, though it remains to be seen how effective those defenses will be at scale.