Tech firms like OpenAI are obstructing a move to ‘democratize’ artificial intelligence (AI) by spinning false narratives, according to Databricks’ CTO Matei Zaharia.
Such narratives can include messaging around doomsday-style risks and exorbitant costs, he said, which are designed to discourage other companies from participating in generative AI development.
Many companies are seeking to make a move in the AI space, whether that be to implement the technology or work to build a generative AI system from scratch.
AI and cyber security
A look ahead to how best practices in AI application capabilities will evolve
“There are definitely the larger providers, like OpenAI, Google, and so on; they have this narrative – and they’re talking in a lot of places about how – first of all, this stuff is super dangerous, not in the sense of a disruptive technology, but even in the sense of ‘it might be evil and whatever’,” Zaharia told ITPro during an interview at Databricks' Data + AI Summit 2023. “It’s very sci-fi.”
“OpenAI – that’s exactly the narrative they’re pushing – but others as well.
“Anytime someone talks about AI alignment or whatever, it’s often from this angle: Watch out, it might be evil. They’re also saying how it’s a huge amount of work to train [models]: It’s super expensive – don’t even try it.
“I’m not sure either of those things are true.”
Zaharia cited MosaicML – the startup Databricks recently acquired for $1.3 billion – as having trained a large language model (LLM) with 30 million parameters that’s competitive with GPT-3, and “probably cost like ten to 20 times less” to train.
While some companies are hedging their bets, there are others that hope the generative AI market comprises just a handful of players, all of which believe only they can understand how to make the technology safe, he added.
The pace of change and innovation, however, defies this notion, with Zaharia impressed with how many smaller models are out there since ChatGPT launched that can rival it.
“A lot of people thought ‘wow it’ll take a long time to catch up’, and now I’ve been surprised – lots of people have been surprised – by the speed at which others have replicated some of that and even done it at a much lower cost.”
In the future, people may be surprised by how general the technology becomes the more that enterprises experiment and find new use cases for LLMs and other generative AI systems, he added.
Get the ITPro. daily newsletter
Receive our latest news, industry updates, featured resources and more. Sign up today to receive our FREE report on AI cyber crime & security - newly updated for 2023.
Keumars Afifi-Sabet is a writer and editor that specialises in public sector, cyber security, and cloud computing. He first joined ITPro as a staff writer in April 2018 and eventually became its Features Editor. Although a regular contributor to other tech sites in the past, these days you will find Keumars on LiveScience, where he runs its Technology section.