Public-private collaboration can help overcome challenges artificial intelligence for social good faces

With many large technology companies supporting the use of artificial intelligence for social good projects and investing in the R&D programs, tremendous progress has been made in this arena. There are a number of projects that leverage artificial intelligence for social good in multiple disciplines—education, transportation, healthcare, and so forth—to leverage the power of artificial intelligence for greater good.

However, as with any other endeavor, there are still many obstacles that need to be overcome while undertaking projects for social good. We list down some of them here:

Data accessibility: Relevant data is needed to arrive at useful insights that can be used to power the social good projects. At present, most of the data is concentrated in the hands of private companies—social media platforms, financial companies, telecom companies, and so forth—and governments that are unwilling to share it. Their concerns are not entirely misplaced as most of the data they possess comprises personally identifiable information of customers, which can be exploited should it reach the bad actors. Data anonymization mechanisms are not entirely fool-proof and prone to hacking.

Useful data: Even if the data were to be made available, sieving out useful data would require a lot of effort and time. Should project owners wish to use tools to automate data cleaning and labeling activities, they would need to invest in additional solutions. This would drive the costs of the projects up, which the available funds or grants may not allow.

Tools and funds: It is still early stages for many social good projects and they need tools and techniques that will lessen the burden on the researchers. Above all, researchers need large amounts of funding and computational resources to support the ongoing efforts.

Bias in AI: There is no denying that bias is a big challenge in artificial intelligence, which can skew the data and propagate prejudices. This inaccurate or biased data can adversely influence the actionable insights, which, in worst cases, can even defeat the purpose of the project. When data is used for social good projects, it is essential that it is free of any bias. Therefore, at present, it is still a mammoth task to ensure fairness of data, especially for social good projects.

Lack of talent: There is a dearth of adequately trained professionals and even businesses face this problem. The number of social good projects is anyway far smaller when compared to the need of skilled professionals in the industry, which further reduces the availability of the right talent.

Lack of principles: It is likely that the authorities and project owners who have access to the AI tools and techniques may misuse them for personal gain or to harm the beneficiaries of the project. This is not to suggest that every project owner would indulge in malpractices, but when there are no established principles to guide project owners, this becomes a possibility which should not be discounted.

The article has been written by Neetu Katyal, Content and Marketing Consultant

She can be reached on LinkedIn.