Let’s not let AI get naughty

As Artificial Intelligence (AI) entrenches itself in our lives, we must keep an eye on it when it strays.

AI is becoming more and more entrenched in our lives, as it enters our everyday lives in unprecedented ways. For instance, AI could help spot early signs of heart failure as well as reveal climate-change tipping points, according to Science Daily. It’s also becoming easier to use AI.

However, much of the world doesn’t trust AI yet. Last week, a British-built robot called Ai-Da that uses cameras and a robotic arm to create abstract art was seized by Egyptian border authorities and detained out of fear that she might be hiding covert spy tools.

Read more: Robot tax: Innovation vs. social adaption of automation

The fact remains that as easy as it is to use and leverage AI, when it comes to training AI, much depends on the intent or ignorance of the trainer. Bias can seep in, creating serious problems.

In September, UN human rights chief, Michelle Bachelet urged that states should place moratoriums on the sale and use of AI until sufficient safeguards are applied. Moreover, she called for AI applications not compliant with international human rights law, to be banned.

As easy as it is to use and leverage AI, when it comes to training AI, much depends on the intent or ignorance of the trainer. Bias can seep in, creating serious problems

“Artificial intelligence can be a force for good, helping societies overcome some of the great challenges of our times. But AI technologies can have negative, even catastrophic, effects if they are used without sufficient regard to how they affect people’s human rights”.

For all the good that AI is doing for us, we must, at all times, keep an eye if it strays from that path.

When AI is Good

Here is some good AI has been doing. As we advance into a more modern and highly connected world, AI is right beside us.

Intello Labs, an AI-powered quality assaying startup, launched Praman a ground-breaking agriculture produce trade exchange platform, with integrated quality assessment. It currently facilitates spot-trading and e-auctioning in agri-commodities.

As an example, its quality assaying technology brings over 95% accuracy to cardamom quality grading as opposed to about 70% using manual methods. This objectivity and precision make fairer pricing and true price discovery possible for both the grower as well as the trader, an unprecedented benefit.

As we advance into a more modern and highly connected world, AI is right beside us

AiStream UK and Delinet India partnered to enable deployment of education, health, and other citizen centric services-at-scale for service providers without the requirement of a fixed or mobile network infrastructure.

In September, Ericsson strengthened its Network Services portfolio with the Intelligent Deployment solution, an agile, digital and modular suite of tools and services that enables communications service providers to roll out, expand, and upgrade networks based on their specific needs and those of their customers.

Artificial intelligence can be a force for good, helping societies overcome some of the great challenges of our times. But AI technologies can have negative, even catastrophic, effects if they are used without sufficient regard to how they affect people’s human rights

Their aim, they said, was a 5G age network that is future-proof and gives quick return on investments.

In the UK, Vodafone is using Ericsson’s Intelligent Deployment solutions to speed up network upgrades. Drones and Lidar-based 3D technology are collecting high-definition imagery and data across 70 different sites to deliver a more digitalised and efficient network deployment.

Startups & AI

The startup ecosystem is definitely leveraging AI to up their game as well as provide unprecedented quality of services.

In September, Appy Pie, a cloud-based no-code app building platform, unveiled the future of chatbots with the launch of an all-new virtual assistant and conversational AI-powered answer chatbot.

The online virtual assistant can be customised as per the need of varied verticals such as banking, healthcare, human resource management, etc.

Gnani.ai, a Samsung Ventures funded company in the voice-led conversational AI automation space announced the launch of its home-grown voice biometrics software. Christened armour365™, the biometric solution boasts of features that cater to new and emerging risks in fraud prevention and information security. It works on 300 plus proprietary audio features and comes with out-of-the-box integrations to multiple contact centre software providers and messaging apps.

Big AI

Big companies have been wowing their audiences with novel usage of AI too.

The recently launched Mahindra XUV700 became India’s first vehicle to integrate Alexa Voice AI. The experience is designed to be an extension of the Alexa experience at home and provides additional automotive-specific functionality, including car control, navigation, and third-party Alexa skills made for on-the-go circumstances, such as finding parking.

The DISH Network Corporation and IBM’s AI-powered automation and network orchestration software and services for new cloud-native 5G network are helping automate US’ first greenfield cloud-native 5G network. The move will bring broad 5G network orchestration to DISH’s business and operations platforms.

IBM’s new Watson Assistant features are designed to help businesses quickly set up a voice agent with IntelePeer. In the last year, clients such as the State of Rhode Island turned to IBM Watson Assistant to address customer problems in the face of rising call volumes.

In September, KPMG released its talent lifecycle management, the AI platform, created in collaboration with dotin Inc, which provides visibility into stages of sourcing, screening, engaging and upskilling talent, aligned to the “Future of Work” theme.

When AI Strays

But time and again, AI strays from its path. In the beginning of this year, Google search for the ugliest language in the world threw up Kannada as the answer, heckling Indian sentiments.

Google apologised as well as explained how the whole thing happened.

Read more: Ecosystem: A look at what AI startups have been up to

“Search isn’t always perfect. Sometimes, the way content is described on the Internet can yield surprising to specific queries,” Google wrote.

To avoid such mistakes, Jordan based Xina AI, the world’s first Arabic interactive voice assistant (IVA) and chatbot, uses AI to solve subtle variances in accents and dialects from one speaker to the next, relying effectively on local linguistic and behavioural preferences to optimise the Arabic speaking customer experience.

An Amoral Power That Can Flip

Apart from all the good AI is doing, it is now also becoming a lie detector to the consternation of those who lie. Reuters reported that certain investors are using AI to scrutinise language patterns and tone, “opening up a new frontier of opportunities to slip up.”

Helpful as it is, AI is an amoral power that can easily be flipped to the advantage of the trainer. Advancing into the future, we must keep an eye

In offices, hospitals, countries, people are feeling more and more comfortable about using AI post-pandemic. But we are also submitting to surveillance at a higher level. Every movement, it seems, can be traced.

Helpful as it is, AI is an amoral power that can easily be flipped to the advantage of the trainer. Advancing into the future, we must keep an eye.