metamorworks - stock.adobe.com

Language models and the metaverse top AI stories of 2021

From moves toward government regulation to the metaverse, language models getting bigger and autonomous vehicle tech slowing, these are some of the biggest stories of the year.

As 2021 closes out and tech moves into the new year, it's time to take stock of some of the big stories and ideas that shaped AI over the year.

A move toward regulation

Early in the year, the federal government made some moves toward regulating AI.

In January, the White House Office of Science and Technology Policy under former president Donald Trump created the National Artificial Intelligence Initiative Office.

The office was included in the National Artificial Intelligence Initiative Act of 2020 and indicated that the government is starting to consider the impact of AI in the business world.

While the slow pace of regulation continued with the new administration, the end of the year brought new signs of activity.

In November, President Joe Biden's chief science advisor, Eric Lander, said during the 2021 Stanford Institute for Human-Centered Artificial Intelligence Fall Conference that he wants an "AI bill of rights," specifying that consumers have a right to explainable AI.

Metaverse takes center stage

An intriguing storyline broke out the during Nvidia GTC Fall 2021 conference, when the AI hardware and software vendor introduced its AI Omniverse Avatar.

The platform generates other AI avatars and is part of Nvidia's Omniverse, a virtual world simulation and collaboration environment.

It is an example of how many vendors are trying to construct the metaverse, a digital world that brings together augmented reality and virtual reality, with 3D avatars, IoT and digital twins.

Other tech giants have touted their vision for the metaverse, including Meta, formerly known as Facebook Inc. Meta says the metaverse will change the internet and how people collaborate in their jobs.

Screenshot of the omniverse avatar
Nvidia's Omniverse avatar, introduced at GTC Fall 2021 conference.

Meta bans facial recognition technology

It's impossible to mention Meta or the metaverse without flagging another big story that broke in November: Meta declared that it will shut down its Face Recognition system on Facebook.

It's impossible to mention Meta or the metaverse without flagging another big story that broke in November. Meta declared that it will shut down its Face Recognition system on Facebook.

As part of the change, people using the facial recognition setting on Facebook will no longer be automatically tagged in photos by the end of the year. The company also said it plans to delete more than a billion people's individual facial recognition templates.

While the move received praise from privacy advocates, it was also met with skepticism. Some observers think the ban applies to only the Facebook uses case, and the company will still widely use the technology.

Large language models

A year after solidifying its partnership with AI research giant OpenAI by revealing that it will exclusively license the powerful GPT-3 language model, Microsoft showed what it plans to do with the license with the preview of its Azure Open AI Service.

The service gives language model customers access to enterprise capabilities such as security, compliance and scale requirements.

Microsoft also entered the competition for which vendor can build the largest language model by partnering with Nvidia to introduce the DeepSpeed and Megatron-powered Megatron-Turing Natural Language Generation Model. Both vendors claim that it is the largest and the most powerful language model so far.

Soon after the birth of the Microsoft-Nvidia alliance, Nvidia competitor SambaNova said it would also provide a GPT AI-powered language model as part of its Dataflow-as-a-Service product for language.

An update on autonomous vehicles

Many autonomous vehicles enthusiasts were convinced that self-driving cars would be commonplace by now.

However, other than a few robotaxis being tested out on the road, fully self-driving cars are still in the future.

Part of the problem is the many hurdles the technology must overcome.

Perhaps chief among them is safety and ensuring that the vehicles can operate without danger in unpredictable conditions. Another challenge is that, perhaps other than robotaxis and the food delivery service, the market for autonomous vehicle is limited.

However, during a pandemic that is still making the future uncertain and dramatically unsettling the present, autonomous vehicles may be finding some place in the economy.

Recently, Walmart and autonomous transportation vendor Gatik sent out driverless box trucks to make deliveries in Arkansas.

This development has some looking at autonomous vehicles to ease the endemic supply chain problems.

Dig Deeper on AI business strategies