AI adaptability to aging is a He Tuber topic that Brainly has been paying close attention to since its establishment. We have reported on many domestic and foreign AI elderly care technologies and trends, and have also followed Baidu, Alibaba, Tencent and other technology companies to visit cases of AI product implementation in nursing homes and home-based elderly care communities.

Witnessing the progress of AI for aging all the way, we strongly feel that in the year 2023, when large language models and general artificial intelligence are running wild, AI for aging has also ushered in drastic changes.

Technology is changing. In Altman's latest speech, he suggested that entrepreneurship and technology development should be based on the realization of general artificial intelligence, suggesting that we may be very close to realizing general artificial intelligence.

People are also changing. The "digital generation" such as those born in the 80s and 90s are beginning to enter middle age and plan for their elderly lives. The "golden generation" such as those born in the 60s and 70s are gradually retiring and have strong needs for utilizing their spare energy, enjoying life, and digital consumption. .

Taking AGI as the premise and human purpose, at this moment, it may be time to re-examine AI elderly care and examine the adaptability of technology to aging.

All the past is a prologue.

Antojito has been continuously reporting on the progress of AI for aging since 2017. At that time, smart elderly care was already a common option for improving the level of smart cities and government governance, and was gradually implemented. Also in 2017, the Ministry of Industry and Information Technology, the Ministry of Civil Affairs, the National Health Commission and others jointly issued the "Smart Healthy Elderly Care Industry Development Action Plan (2017-2020)", which clarified the definition of smart healthy elderly care.

Looking back at the age-appropriate AI at that time, there are several main features:

The implementation of AI products among the elderly is mostly welfare and inclusive. For the government, community care for the elderly in the jurisdiction and care services in nursing homes are all welfare investments. For technology companies, aging-friendly projects are also regarded as social responsibility and public welfare projects. The core is "smart assistance to the elderly" and reduce the pressure and burden caused by elderly care.

The core technology of smart elderly care solutions in 2017 and before is the Internet of Things + big data, which uses video terminals, electrical appliances, induction monitoring equipment, etc. to provide care for home-based elderly care and elderly people living alone in empty nests. At that time, the AI ​​products that could really be applied on a large scale focused on "companionship", mainly smart speakers, which reduced loneliness by chatting with the elderly and dispatched some simple home appliances.

A Beijing nursing home we visited built an "intelligent Internet-connected nursing home model room" that uses smart speakers to turn off lights, turn on air conditioners and fans, and other operations by voice, eliminating the need for the elderly to manually control the room.

As for some more complex embodied smart products, such as "robotic bear" care assistants that can hold the elderly and robots that dress the elderly, to be honest, they were limited by technical reliability and cost and were still considered "black technology" at the time. category, it has not been widely popularized in China.

It can be said that AI was suitable for aging at that time. Whether it was practitioners such as governments and enterprises, or reporters and observers like us, they all belonged to the "niche". After all, there was a lack of conditions on either the demand side or the supply side. There are not that many elderly people, and they generally have low awareness of smart technology, leaving a lot of gaps in demand. Moreover, AI based on traditional machine learning and small models has limited scenarios in which it can exert value.

So, has the turning point of change come?

The most critical factor in promoting the aging-friendly upgrade of AI should be that the composition of the elderly group has changed.

Around 2020, the golden generation began to gradually retire, and those born in the 1980s entered middle age. The generation born between 1960 and 1970, known as the golden generation, became the backbone of social development and economic growth around 2010. With the rapid growth of China's economy, as well as factors such as environmental improvement, public security optimization, and improvement of public medical standards, Whether it is physical condition or wealth accumulation, the elderly have better conditions than before.

The "2023 China Private Wealth Report" released by China Merchants Bank shows that more than 51% of high-net-worth individuals are 40 years old and above, and more than 90% are 30 years old and above.

As the golden generation ages, its impact on the silver economy is very obvious.

At the 2023 Elderly Expo, many grandparents walked around the exhibition by themselves in wheelchairs and crutches, actively looking for their favorite elderly care and assistance products, instead of blindly relying on younger generations or traditional equipment suppliers.

Needless to say, those born in the 1980s have higher digital literacy than previous generations, and have a proactive awareness of retirement planning. They are also young, have an international perspective, and are very concerned about cutting-edge retirement technology.

The elderly group and its reserves have been transformed into a group of people with demand, understanding and consumption ability for AI products, which has laid the foundation for the upgrade of AI for aging.

The explosion of large models in 2023 will bring a burst of east wind to the adaptive aging of AI.

So, what practical progress will be made in AI aging-friendly products in 2023?

Let’s talk about the ones that have been implemented first.

First, changes in human-computer interaction:

In meeting the needs of the elderly, traditional AI may still be "intelligent for the sake of intelligence". The diverse and refined functions are difficult for the elderly to master, and the designs, interaction methods and quality of different manufacturers vary. Similarly, it is easy for the elderly to have problems such as inability to use, difficulty in using, difficulty in operating, and potential safety hazards.

A developer from a mobile phone manufacturer told me that during user interviews, one of the biggest problems for elderly users is that when they receive a customer service call, they are prompted to press the XX button. The elderly understand it, but their hand movements are slow and the smartphone touches the phone. The screen is very sensitive, so it is easy to accidentally touch it, resulting in repeated failed operations.

The actual decline in physical functions of the elderly and the lack of targeted adjustments in smart products for aging have led to many elderly users becoming repulsive and unwilling to deal with smart machines.

The large language model has powerful understanding, analysis and interaction capabilities, truly realizing a natural language user interface (LUI, Language as User Interface). It also supports multi-modal recognition and free conversion between videos, pictures, speech and text. It can subversively change the way the elderly interact with smart machines, which is very helpful for non-digital native elderly people to use AI products.

For example, complex steps that originally required manual operation can now be described in a sentence of natural language; prescriptions that originally required typing by hand can be automatically recognized into text by taking a photo, making it easier for the elderly to communicate with doctors and nurses online.

It can be said that the large language model’s mastery of human knowledge, unprecedented understanding and generalization capabilities can effectively bridge the digital divide and allow elderly people who are unfamiliar with smart products to easily overcome difficulties.

In 2023, almost all leading domestic mobile phone manufacturers will upgrade their mobile operating systems with large model capabilities. Among them, vivo’s OriginOS 4 particularly emphasizes the new simple mode, elder care and elderly color research and other aging-friendly functions. OPPO's ColorOS 14 has added many new smart integration experiences. For example, when you encounter a problem using the phone, you don't need to look up the solution yourself, you can directly ask the new Xiaobu Assistant. The call transcription function uses large models to automatically extract key information to avoid missing key information. , are very friendly to the elderly group.

In addition to the above C-side scenarios, human-computer interaction supported by large models has also brought changes to the smart elderly care services of the government and public sectors. For example, the smart elderly care in Hexi District of Tianjin is supported by JD.com’s large-scale Yunyanxi model. Through intelligent voice outbound calls, it serves 200,000 elderly people in Hexi District for home care. The platform will make daily calls to inquire about their physical conditions and chat. This greatly reduces the work difficulty of community personnel.

Another application that has already been implemented is the advancement of intelligent health management.