ChatGPT co-creator wants government oversight on tech she helped build

The trend: Calls for AI regulation are getting louder as the technology’s use expands.

• OpenAI CTO Mira Murati, who helped build ChatGPT, acknowledged that the chatbot isn’t always accurate and also called for government oversight of the technology.
• EU industry chief Thierry Breton is pushing for the passage of strict rules governing AI under the proposed AI Act.
• Executive director of global at Deloitte AI Institute Beena Ammanath has warned that a generative AI arms race, like the one between Microsoft and Google, could have “unintended consequences.”
• Ammanath also said that companies’ cavalier AI deployment is like “building Jurassic Park, putting some danger signs on fences, but leaving all the gates open.”

Use only as directed: OpenAI co-founder Sam Altman warned after the chatbot’s release last year that “it’s a mistake to be relying on it for anything important right now.” The problem is that the directive has been widely ignored.

• Last month, a judge in Colombia used ChatGPT to inform a ruling on a case involving health insurance coverage for a child with autism.
• The chatbot has an IQ of 83, well below the US average of 97.43, according to the Ulster Institute for Social Research, and is being used to write scientific reports, craft legislation, and write code, among other uses in business and education.
• Users are circumventing OpenAI’s safeguards by getting the bot to say positive things about drug abuse and give advice for how to smuggle drugs into Europe.

Moving fast and breaking things? AI chatbots don’t think. Instead, their algorithms enable them to string together information scraped from the web in creative ways. They’re generating interesting content that’s also laden with inaccuracies and bias.

• The risk of AI fueling the misinformation problem posed by social media is concerning, but the tech comes with bigger problems, like copyright infringement, more sophisticated cyberattacks, undermining education, and disrupting human occupations.
• Tech companies are aware of these risks, but some are under pressure to please investors and beat the competition in spite of them.
• The generative AI adoption frenzy is being fueled by figures like ARK Invest’s Cathie Wood, who says the tech will spur “the most massive productivity increase in history.”
• 51% of 113 EU-based AI startups surveyed by appliedAI say the AI Act will slow their AI development work.
• Slowing development could give companies time to build AI systems that are more helpful than current models and equipped with foolproof safety features.
• With Keen Technologies founder John Carmack predicting that artificial general intelligence (AGI) will be achieved in the 2030s, the clock is ticking for regulators to address AI’s existential risk for humanity.