Three major challenges facing the popularization and application of artificial intelligence

The use of connected smart devices is growing rapidly, but they are not yet ubiquitous, and widespread adoption of artificial intelligence faces many challenges.

Many friends may have realized that artificial intelligence has affected our lives. Achieving ubiquitous artificial intelligence could be life-changing in the future. For example, our food never goes bad and everything we eat is healthy. Even out shopping, the store will immediately know that we have walked in and start recommending customized products, which is why it is crucial to understand and break down barriers to AI adoption.

3 obstacles to the adoption of artificial intelligence applications:

• Hardware and hardware compatibility are key, and the technology is not yet mature.
• People have reason to worry about privacy issues.
• The necessary technology is very expensive.

The first challenge: Hardware

For example, a smart refrigerator. Samsung launched such a device in 2018, but it was more of a novelty. According to data from an organization, smart refrigerators were the most popular in North America in 2019, accounting for 31% of the global market. However, these are not Star Trek-style devices with touch screens, instead they are smart thanks to internal circuitry that allows for greater efficiency and self-monitoring, and users may not even realize how smart the device is.

Major appliances take longer to replace than mobile phones, and consumers will replace larger items as needed, so they are less likely to buy a new phone simply because it is more efficient than because it is more efficient. Possibility of buying a new phone with slightly better battery life.

Updating hardware is not a simple task either. We can't just add a Wi-Fi card to any refrigerator and hope it broadcasts service records to the local repair center. Most of the electronics in our lives are not modular or designed to scale well beyond their current design. This is a serious limitation for integrating IoT devices, because as anyone with just a few smart light bulbs will tell you, hardware and hardware compatibility is key, and we're not there yet.

That said, smart refrigerators, as well as other examples of smart hardware, are absolutely necessary to build a future life driven by artificial intelligence. It will take a while for us to agree on a future of AI-driven hardware that some of us will buy into sooner than others. This group of early adopters is critical to leading the way to mass adoption, helping to address vulnerabilities and proving that these products not only work, but bring value to people’s lives.

Second Challenge: Privacy

We are entering an era where personal data is more valuable than ever, and so are consumers Become aware of this fact. A 2019 report showed that more than 60% of respondents believed connected devices were creepy, which could reduce adoption of such devices.

While all of this may sound daunting, there are some interesting innovations that address these pain points. We may be enjoying the benefits of this kind of thinking without even realizing it. To understand this, we have to enter a room full of network devices.

Most of us are familiar with server rooms, thanks to some of the mundane yet high-tech data centers we see in TV shows and movies. What most consumers don't realize is that businesses don't upgrade all their data center hardware at once. Just like when you buy a new laptop, you probably don't buy a new router, over time data center components get replaced here and there, and it can end up becoming a patchwork of vendors and services.

Some time ago, network administrators unified management while allowing the underlying system to micromanage individual components. This requires special software to integrate all the different requirements of all the different devices, controlling them as needed while obfuscating the details for managers.

As data centers continue to be upgraded, more and more privacy is embedded in them. While we can't yet fully trust every step we take toward artificial intelligence, we can expect that in the next few years, most data centers will be privacy-centric.

Final Challenge: Cost

The cost of current AI solutions is often prohibitive. However, this is not always the case.

We’re already pushing AI to the edge in a more cost-effective way, layering software on top of existing hardware rather than waiting for dedicated AI-specific chips . We can add functionality to machines by leveraging their networks and power grids.

Going back to our not-so-smart refrigerator, what will happen if the electric box is replaced by a smart electric box that detects the refrigerator in your home based on its power usage? ? The smart electric box will know the manufacturer and model of the refrigerator and make decisions about the contents of the refrigerator based on this information. Add a smart kitchen camera, or an anti-embedded scale, and we can add sensors without adding much cost.

Ultimately, the best AI solutions will cross all these barriers. They will push AI to end consumers without relying on specialized chips, which will require consumers to replace their devices with new ones. After all, ubiquitous AI relies on it wherever it is needed. ​

The above is the detailed content of Three major challenges facing the popularization and application of artificial intelligence. For more information, please follow other related articles on the PHP Chinese website!