Google Tries To Straddle Being 'Bold' And 'Responsible' In AI Product Blitz

Google unveiled new AI products at its annual developer conference.

After months of criticism that it has fallen behind more nimble rivals in the race to mainstream and monetize artificial intelligence, Google showed off an array of new AI tools it hopes will help it regain the "AI first" pole position CEO Sundar Pichai boasted about seven years ago.

At Google's annual I/O developer conference on Wednesday, the company unveiled a barrage of AI products: an editing tool that will fix your photos even if you accidentally cut something out of the frame, a tool that will write your emails and elaborate on certain themes, and a revamped search page that puts AI-generated results front and center.

The onslaught of announcements come as Google has been caught in a seemingly impossible situation. The tech giant has faced criticism that it has moved too slowly in releasing AI products, allowing rivals to catch up in a field Google helped pioneer. Meanwhile, critics worry about the potential harm of AI, unleashing the technology into the world too quickly, without fully understanding the consequences.

As Google tries to repair the image that it's lost its drive and ability to innovate, it's looking to have the best of both worlds. The company's new mantra is to be "bold and responsible" when it comes to AI — a phrase that company executives repeated over and over again during a 2-hour keynote event at the Shoreline Amphitheater in Mountain View, California.

The company knows that understanding the approach may require some mental gymnastics, so it attempted to explain the logic. "While there's a natural tension between the two, we believe it's not only possible, but in fact critical, to embrace that tension productively," James Manyika, Google's senior vice president of technology and society, said on stage. "The only way to be truly bold in the long term is to be responsible from the start."

At the event, Google tried to put both "bold" and "responsible" on display, introducing products that stretch the bounds of reality and others that try to preserve it. A new photo editing feature lets people touch up pictures by not only cosmetically fixing things like lighting, but by altering elements in the photo. For example, if you took a photo with a waterfall, you could use AI to make it look like water is splashing on your hands. Or you could alter the sky to make it look like it was sunny when it wasn't.

On the "responsible" front, the company debuted tools intended to curb AI-generated misinformation. Google announced "watermarking" and metadata features to identify if content was created by AI. People will also be able to self-label photos as AI-generated. The issue came to light in a major way in March, when an AI-made image of Pope Francis in a puffy jacket went viral and fooled the masses.

Google, perhaps more than any other company, is in a precarious position when it comes to moving too fast or slow. The company, which has invested heavily in AI for years, has been struggling to regain its footing since late last year, when OpenAI released ChatGPT and touched off a battle in consumer AI. The product's success blindsided Google: ChatGPT had been built using a technological breakthrough developed by Google researchers, and Google had announced but not released a similar chatbot named LaMDA two years earlier.

Google's sluggishness in releasing AI products threw the company into an existential tailspin. Employees, investors and industry observers began to question Pichai's leadership, deriding him as slow and indecisive. Around Google, cofounders Larry Page and Sergey Brin re-entered the fold, even tinkering with code after being hands-off for years. Pichai reportedly upended the plans of several teams to refocus them on AI products, with the aim of announcing them at Wednesday's keynote.

Other announcements include a new AI model called PaLM 2 that's even more powerful than the last, new Pixel phones chock full of AI features, and a program called Labs to let users test new AI products.

But with every new announcement, Google tried to strike a sober tone. It was a contrast to Google's attitude at I/O five years ago, when the company unveiled Duplex, an eerily human-sounding AI that was meant to automate appointment bookings. The demo both impressed and disturbed the public, drawing immediate backlash as people began to debate the ethics of a robot masquerading as a reservation-seeking human..

Wednesday's event steered clear of such bombast. "We must also acknowledge that it's an emerging technology that is still being developed," Manyika said. "And there's still so much more to do."