Ethical AI - What is it? How to go forward?

Can we start from stone age? There was a time when sharpening stone was a technology. It was a leveraging point through which hunting was much easier. Human civilisation was curious. It has invented hundreds of technologies since lakhs of years. At one point oil and coal became a leveraging point. Those created industrial revolution in several countries like UK, Germany, France etc over last few centuries. What is the leveraging point now? Is it oil, coal? Yes, they are of course a leveraging point still, but it is not enough to create a new industrial revolution. The new leveraging point is 'curiosity' or 'ideas'. This curiosity generated 'artificial intelligence' which is one of the biggest sources of productivity. To utilise AI, we need data to feed into the system. Reliance Industries owner Mukesh Ambani rightly said 'data is the new oil'. Companies are using AI to build scalable solutions. But is it ok to use data, however we want? Let's discuss this.

Once company use AI to build a large business, it not only increases their business, but it also increases their brand value and legal risks. Regulators investigated Goldman Sacks for their AI algorithm which discriminated male over female and granted larger credits to men. Everyone knows how Meta created a bad reputation over Cambridge Analytica . So, it is immensely important to have a data ethics for every company which are trying to get leverage from AI. Amazon has burnt out several years to build an AI powered hiring software. But it has not found any model on how not to discriminate against women in hiring. Sidewalk labs , a subsidiary of Google lost 2 years of work and $50 million due to lack of its data ethics. It tried to build an IOT powered smart city inside Toronto where they get strong opposition by local citizens and local government officials due to lack of data ethics standard. These are some risks which business need to investigate.

What should be the way forward? Here are few steps which can be taken -

Data Ethics Standard for Organisation:

HubSpot founder and CTO Dharmesh Shah has created a slide deck for their employees. He believed culture is a product which can be built over iterations just like another product. The organisation believed in their culture, and it is transmitted to every employee. The same thing needs to be done for every company. They need to fix their data ethics and make sure it has created a culture in the company. But who is going to fix the standard of data ethics? It is a matter of morality. Morality differs person to person. Every person has their own personal moral code. It is a difficult problem to solve. However, new age founders are not only strong in their business areas, but they also have strong viewpoint on philosophy. They are well equipped to build a strong data ethics inside organisation. However, a discussion with academic expertise (eg. philosophical experts, epistemologist etc) also could be helpful.

Health Industry Success:

Health Industry had so many violations earlier days. There are many stories on how human body (opponent soldier) was cut alive to improve on medical science, during second world war. Since last 50 years the industry has improved heavily. With the invention of CRISPR technology , it has developed a concern again. But still, it is heavily in a decent shape. Expertise from this industry could be taken to build a good data ethics for entire IT industry.

Incentive for employees:

The entire cavillation is run by incentive. Media shows negative aspect of life as it has a better incentive for them. Entrepreneurs are mostly positive because their incentives are heavily tied with that. Company may have good data ethics standard but unless and until employees has a strong financial incentive it won't work. Big tech companies provide bounty if someone raises a security breach in their website or mobile app. The same must be done for data ethics as well. Employees need to be awarded with financial help if they found any data ethics breach, which could happen. Anyway, it will save the company months of time, monetary loss, and legal issues.

We are in a tricky situation. A strong data ethics will lead to slow progress for a company. But slow world is fine when our data is at risk. Microsoft , Google has prepared their principal years ago. But is that enough? What should be the solution? What's your views on this?