Tim Cook says A.I. companies will also need to regulate themselves

Apple CEO Tim Cook believes companies should modulate themselves erstwhile it comes to artificial intelligence.

In an interrogation that aired Tuesday, helium told ABC's "Good Morning America" that ample connection models – the AI tools that powerfulness chatbots similar OpenAI's ChatGPT and Google's Bard – amusement "great promise" but besides the imaginable for "things similar bias, things similar misinformation, possibly worse successful immoderate cases."

Cook besides offered his thoughts connected regularisation and guardrails, saying that they're needed, but that AI is almighty and the tech's improvement is moving quickly. 

"If you look down the road, past it's truthful almighty that companies person to employment their ain ethical decisions," Cook said. "Regulation volition person a hard clip staying adjacent with the advancement connected this due to the fact that it's moving truthful quickly. So I deliberation it's incumbent connected companies arsenic good to modulate themselves." 

Cook's comments connected AI's fast-moving improvement and sizeable powerfulness travel a high-profile connection past week astir AI's risks, signed by tech leaders similar OpenAI CEO Sam Altman, DeepMind CEO Demis Hassabis and Anthropic CEO Dario Amodei. But AI researchers and critics increasingly say that overarching statements astir AI's increasing powerfulness and threats distract from the real-life harms that the algorithms tin origin to marginalized communities close now, alternatively than successful an unspecified future. 

In the interview, Cook said that portion Apple intelligibly integrates AI into its products today, the wide nationalist doesn't needfully deliberation astir those features arsenic AI, aboriginal adding that he's excited astir ChatGPT's applications and that the institution is looking astatine the instrumentality closely. 

His statements came one time aft the keynote speech of Apple's yearly developer conference, WWDC, wherever the institution touted caller applications of instrumentality learning for the iPhone, iPad and Mac, specified arsenic AI models that tin issue astute prompts to users for imaginable diary entries, amended AI tools for autocorrect and dictation successful messaging and much broad facial designation features for identifying friends, household and pets successful photos. 

Watch the afloat interrogation connected GMA.