Technology, media and telecoms (TMT) regulators will ponder rather than act on AI. Increased use of AI to generate deepfakes in the US presidential campaign may be the catalyst for substantive regulation. Debate regarding access to, and ownership of, data will continue with little regulatory change. For many industries, the focus will remain on operational efficiency. AI-based virtual assistants will gain significant traction. Smart speakers will become gateways to automated homes and AI-based automation.
Listed below are the leading TMT trends in artificial intelligence, as identified by GlobalData.
Where the pursuit of profits collides with fear, hype, and loathing
The artificial intelligence world is marked by contrasts, and there are parallel but separate movements that will gain momentum in 2020. Technology marketers will continue to falsely claim their products are AI-powered, while data scientists will create incredible new AI applications. There won’t be an end to the hype in 2020, but there will be a continuing flow of new product releases.
The biggest revenues will be earned through AI-enabled business process improvements
Machine learning and deep learning will be the primary commercial uses of AI in 2020. Machine learning tools will increasingly inform companies on business process efficiency across the enterprise. The accuracy of existing AI programs will improve as algorithms become better at identifying false positives and refine recommended actions. More use cases will continue to develop in call centre optimisation, operational efficiency, cognitive process automation, smarter software-defined networking, and predictive or prescriptive maintenance.
Computer vision will be the next big thing
There are many applications of computer vision. Security applications gain the most media attention, given the ethical implications of greater citizen surveillance. However, there are many enterprise applications, particularly around improving visual inspection. Video recognition will help protect against copyright infringements and detect deepfake videos. Computer vision can also be used to identify early stages of cancer.
Bias will become a bigger problem
One of the biggest challenges AI faces relates to the information fed into learning algorithms. The world is awash with examples of unintended (and sometimes intentional) consequences – racism, sexism, and psychopathy – caused by an algorithm’s training data. The use of deep learning and other unsupervised techniques will exacerbate the criticism of bias.
Hype will get worse and understanding won’t improve
AI is a tech marketers’ dream. Its loose definition means it’s easy to badge a technology product as ‘AI-powered’ without falling foul of trading standards. Almost all technology is badged as AI while, in practice, very little is.
This is an edited extract from the TMT Predictions 2020 – Thematic Research report produced by GlobalData Thematic Research.