As the tech world embraces ChatGPT and other generative AI programs, the industry’s long-standing promises to deploy AI responsibly could quickly be overwhelmed by competitive pressures.
Why is this important: Once again, tech leaders are playing the game of “build fast and ask questions later” with a new technology that has the potential to cause profound changes in society.
- Social media began two decades ago with a similar market rush. First came the excitement – later the damage and regret.
Catch up fast: As machine learning and related AI techniques have sprung up in labs over the past decade, academics and critics have sounded the alarm about the potential harm the technology could foster, including misinformation, bias, hate speech and harassment, loss of confidentiality and fraud.
- In response, companies have made reassuring statements about their commitment to ethical reviews and screening for bias.
- High-profile missteps — like Microsoft Research’s Twitterbot “Tay” in 2016, which was easily tricked into repeating offensive and racist statements — have made tech giants reluctant to push their most advanced AI pilots into the world.
Yes, but: Small businesses and startups have much less risk, both financially and reputationally.
- This explains why it was OpenAI – a relatively small competitor in the field – rather than Google or Meta that started the current generative AI frenzy with the release of ChatGPT late last year.
- Both companies have announced several generative AI research projects, and many observers believe they have developed tools in-house that meet or exceed ChatGPT’s capabilities, but have withheld them for fear of infringement or liability.
ChatGPT “is nothing revolutionary” and other companies have matched it, Yann LeCun, chief scientist at Meta AI, said recently.
- In September, Meta announced its Make-a-Video tool, which generates videos from text prompts. And in November, the company released a demo of a generative AI for scientific research called Galactica.
- But Meta brought down Galactica after three days of scathing criticism from academics that it generated unreliable information.
And after: Everything restraint giants like Google and Meta have shown to date may now be eroding as they seek to demonstrate that they haven’t fallen behind.
How it works: The dynamics of startup capitalism and Silicon Valley techno-optimism create powerful incentives for companies to bring new products to market first and worry about their social impact later.
- In the AI image generator market, OpenAI’s popular Dall-E 2 program came with built-in safeguards to try to prevent abuse. But then a smaller rival, Stable Diffusion, came along and stole Dall-E’s thunder by offering a similar service with far fewer limitations.
- Meanwhile, the US government’s slow pace and limited ability to produce laws means it rarely stays ahead of new technologies. In the case of AI, the government is almost entirely in the “making voluntary recommendations” stage at this time.
Be smart: Technology leaders are haunted by the idea of the “innovator’s dilemma”, first articulated by Clayton Christensen in the 1990s.
- The innovator’s dilemma says that companies lose the ability to innovate once they become too successful. Incumbents are required to protect their existing businesses, but this leaves them vulnerable to new, more nimble competitors with less to lose.
Our thought bubble: The Innovator’s Dilemma has accurately mapped out how the technology business has operated for decades. But the AI debate is more than a business issue. The risks could be national or planetary, and humanity itself has much to lose.