Page 749 - AI for Good Innovate for Impact
P. 749

AI for Good Innovate for Impact



               impairments) and multilingual communities. On-device AI capabilities, such as real-time
               speech-to-text, image recognition description, and sign language recognition, help them
               interact with the world more conveniently, participate in social activities, and access information
               and resources.  Due to the widespread use of smartphones, the features can benefit a broader         4.9: Accessibility
               user group, narrowing the digital divide, fostering social inclusivity, and reducing inequalities
               caused by disabilities or impairments.

               Responsible Consumption and Production

               The on-device AI approach reduces dependence on cloud servers, minimizing the energy
               consumption required for data transmission and processing, aligning with the green and
               environmentally friendly principles. This low-energy, high-efficiency computing method
               reduces carbon emissions and supports sustainable development. Moreover, protecting user
               privacy and preventing data abuse reflects the commitment to consumer rights. By providing
               users with sustainable products and services, we advocate for more environmentally friendly
               and responsible consumption model.

               Use Case Status: The use case is part of a larger product development

               Partners: China Foundation for Disabled Persons(CFDP) [9], China Association of Persons with
               Visual Disabilities (CAPVD), Audiology Development Foundation of China(ADFC)

               2�2     Benefits of the use case

               Reduced Inequality: By offering AI-assisted features directly on mobile devices, the use
               case enhances access to technology for underserved and marginalized groups, including
               individuals with visual or hearing impairments and speakers of diverse languages. On-
               device AI functionalities—such as real-time speech-to-text conversion, image recognition with
               descriptive output, and sign language recognition—empower users to engage more easily
               with their environment, participate in social life, and access critical information. Leveraging
               the widespread availability of smartphones, these features reach a broader population, help
               bridge the digital divide, and promote greater social inclusion by addressing barriers caused
               by disability or language limitations.

               Responsible Consumption and Production: The use of on-device AI reduces reliance on cloud-
               based computing, thereby lowering the energy consumption associated with data transmission
               and remote processing. This approach promotes environmental sustainability through energy-
               efficient computing, ultimately contributing to lower carbon emissions. Additionally, by enabling
               local data processing, the system upholds user privacy and helps prevent misuse of personal
               data. These measures support a more responsible and sustainable technology consumption
               model that aligns with environmentally friendly and user-respecting design principles.


               2�3     Future Work

               Enhance the application effects of on-device AI in empowering disabled individuals and smart
               scenario and accelerate the process of sustainable development internationalization.

               1.    Technology Optimization and Product Innovation

               Optimize on-device AI algorithms, focusing on improving model lightweighting and efficiency,
               adapting to mobile phones with different hardware configurations. Integrate the latest



                                                                                                    713
   744   745   746   747   748   749   750   751   752   753   754