The European Union finalised a set of rules last week to regulate how artificial intelligence (AI) will be used. The objective of such regulation is to provide protections for the public while giving clarity for tech developers on how to proceed with the technology. The EU regulation, among other things, bans the use of biometric systems that will identify people based on race and sexual orientation, bans predictive policing systems, includes transparency requirements, provides protections for copyright holders, and bans emotion recognition in the workplace. China had also issued some regulations earlier. Other countries are bound to follow.

AI is like derivatives and cryptocurrency. We should not dismiss it as complex and try and understand what it’s all about for otherwise it will come back to hurt us. When decision-making algorithms incorporating AI are built into products and services we use, it will certainly impact all of us. Since it is hidden, we would very soon not know how AI is impacting all that we do and that is scary. Supposedly, the potential use/misuse of AI is one of the issues on which the board of the company OpenAI considered seriously in its decision to remove Sam Altman from his CEO role just a few weeks ago.

Many firms have been profitably incorporating AI into their systems for some time now. For instance, Amazon.com has significantly automated its warehouse operations in how goods are picked and placed into bins for shipping. At this stage, the items go through an imaging station to confirm if the right item is being shipped. By incorporating AI here, the station compares the item with images of damaged and undamaged items, learns about different damages over time, and prevents a damaged item from being shipped. Compared with the manual alternative, the company reports it has been three times more effective in reducing shipping damaged items.

In contrast, AI is also being built into hospital systems that monitor patient vitals and issue alerts for action. In one case reported recently, the system did not issue an alert when a patient had an elevated white blood count based on records of previous patients with similar combination of problems. However, the supervising nurse who knew that this case was different because of the patient’s history could not overrule the system’s decision and take action under the hospital’s rules.

Rapidly transformative

AI clearly can be a benefit and a curse. Venture capitalist Vinod Khosla, one of the investors in OpenAI, says that in about 10 years, 80 per cent of doctors and accountants may potentially be replaced due to the use of AI systems. That is surely rapidly transformative. AI will boost productivity and growth of economies as it is incorporated in more and more sectors. What do we do with the people being replaced? One charitable suggestion is that part of the gains can be used to fund more extensive universal basic income programmes so we avoid hunger and give people a chance to think of what they would like to do next. Employment agencies need to become more than application filing centres and become counsellors.

AI has already begun to be used to spread misinformation with generated images and text that seem realistic. Who do we believe now when we log into social media? Fortunately, AI-based programmes can also be used to distinguish what has been generated by AI and that is how educational institutions will have to deal with the homework their crafty students submit.

But another subtle problem was pointed out by Taiwanese AI expert Kai-Fu Lee. The generative AI models being developed in the West are all being trained in English on American and British data. That is where the giant data repositories are available to develop the models. And thus, these decision-making routines will reflect the values and beliefs of those societies and may not suit other countries. If you are smarting how ‘pandit’ has become a derogatory term in the West, watch out.

Already, much knowledge creation in the world takes place in English and the spread of AI can make the world more culturally and socially parochial. Even though their resources may be limited, other countries need to jump on the bandwagon right now.

The writer is an emeritus professor at Suffolk University, Boston

comment COMMENT NOW