Meta technology chief defends tech titan’s AI strategy

Meta’s chief technical officer is quick to refute claims that the business has lagged behind rivals like ChatGPT in the rapid rise of generative AI in the tech sector.

Meta chief defends tech: At the company’s latest Connect developer conference, Andrew “Boz” Bosworth stated to AFP that “the majority of the world’s population will have their first experience of generative artificial intelligence with us.”

During the event, Meta demonstrated tools for creating photos or textual content through spoken commands, and they also showcased chatbots infused with AI that possess personalities.

Competitors like Microsoft and Google, which have released generative AI products and made significant investments in the field, have compared Facebook and Instagram’s parent firm to, suggesting that it has lagged behind.

Bosworth insisted that Meta is keeping up by integrating AI throughout its global platforms ever since ChatGPT went live late last year.

Regarding the buzz surrounding AI products from other businesses, Bosworth remarked, “There are lots of cool tools, like Stable Diffusion, for generating images.”

The CEO quickly clarified, however, that many of these new platforms require practise and advanced computer knowledge to understand.

Bosworth noted that Meta’s technology enables customers to easily request an image of “hedgehog on a bike” or “happy birthday to a marathon runner,” for example. “We wanted the results to be great and fast, even on smartphones,” Bosworth added.

Real or fake news?

When it comes to generative AI, Meta has decided to take a more conservative approach for the time being.

Meta launched a generative AI chatbot named “Galactica” two weeks before the launch of ChatGPT in November 2022. It was focused on scientific research.

Galactica could write essays and do maths, but she occasionally made up solutions.

Meta swiftly eliminated the tool, which, according to Bosworth, was a mistake, as reported by AFP.

Bosworth declared, “If it had been up to me, I would have left it.” We had advised users to approach with caution because our chatbot was “capable of saying anything,” according to the statement.

While this was going on, Meta’s competitors were still offering AI products despite the possibility of strange results known as “hallucinations,” which made the company in question appear foolish.

However, Meta will probably err on the side of caution when it comes to establishing safety criteria for its own inventions after years of criticism around content moderation at its biggest social media platforms throughout the world.

While the other industry titans released their AI solutions, Meta upgraded its own AI model and released Llama 2 earlier this year as open source, allowing programmers to modify it to make their own chatbots.

Metaverse

Bosworth, a Facebook employee since 2006, most recently served as the executive in charge of the newly renamed Meta tech company’s augmented and virtual reality innovation section.

It was said that changing Facebook’s name in 2021 represented Zuckerberg’s belief that the metaverse will be the next significant computing platform.

The move, according to critics, was actually a part of a plan to improve Facebook’s reputation following charges that it prioritised profits over user safety and wellbeing.

Furthermore, despite spending billions of dollars on its metaverse goal, Meta is still a long way from realising it.

Bosworth acknowledged that, up until recently, adoption of Horizon Worlds, Meta’s immersive social networking platform, has lagged behind expectations.

Bosworth made the sarcastic remark, “We should have had legs sooner,” in reference to giving avatars in virtual worlds limbs.

At Connect, Meta mostly avoided discussing the metaverse in favour of highlighting goods like Ray-Ban smart glasses, which let users broadcast what they view live.

The usage of “mixed reality” devices like its VR headgear overlays digital content on the environment rather than entirely submerging the user.

The new Quest 3 headsets, which go on sale in October, now have the option to switch from virtual reality to augmented reality.

A Quest user’s eyes won’t be seen to others.

Bosworth noted that making the headgear incurs costs and remarked, “We’ve tried it, and the result can be quite off-putting.”

The IT industry is eagerly anticipating the $3,500 Apple Vision Pro, which will launch early next year and will cost more than the $500 Quest 3.

Nothing in Apple’s high-end offering, according to Bosworth, prevents us from building it.

However, Meta never imagined that spending so much money to construct anything “was going to help our developers reach a big enough audience to be meaningful.”