Verdict lists five of the most popular tweets on artificial intelligence (AI) in Q3 2021 based on data from GlobalData’s Influencer Platform.

The top tweets were chosen from influencers as tracked by GlobalData’s Influencer Platform, which is based on a scientific process that works on pre-defined parameters. Influencers are selected after a deep analysis of the influencer’s relevance, network strength, engagement, and leading discussions on new and emerging trends.

Related Company Profiles

View all

The most popular tweets on AI in Q3 2021: Top five

1. Dr Kate Crawford’s tweet on AI tools failing to identify Covid-19 cases

Kate Crawford, co-founder and director of the AI Now Institute at the New York University (NYU), shared an article on how hundreds of predictive tools developed during the pandemic to detect the SARS-CoV-2 virus failed. Some of these tools even raised concerns about being harmful as they were not tested and could have led to wrong diagnosis. A report from the Turing Institute, UK’s national centre for data science and AI, concluded that AI tools had little impact in diagnosing and triaging Covid-19 patients.

The findings of Turing Institute’s report were similar to the results from two major studies including one conducted by Laure Wynants, an epidemiologist at the Maastricht University in the Netherlands, and her colleagues. The study analysed 232 patient algorithms to diagnose or understand the severity of the disease. The analysis found that the predictive tools were of no clinical use while just two of them were found to be effective for future testing, the article highlighted.

Another study led by Derek Driggs, a machine learning researcher from the University of Cambridge, and his team analysed 415 published deep learning models and found them to be ineffective in diagnosing Covid-19. The tools were also unable to predict the risks of exposure to medical imaging techniques such as chest computer tomography (CT) and chest x-rays, the article detailed.

Username: Dr Kate Crawford

Twitter handle: @katecrawford

Likes: 236

Retweets: 116

2. Vala Afshar’s tweet on deep learning AI being top disruptive technology in 2021

Vala Afshar, chief digital evangelist at Salesforce, a software company, shared an article on 15 big tech innovations of 2021, according to a report by ARK Invest, an asset management company. The report highlighted that deep learning AI was the top disruptive technology this year and is expected to add $30 trillion to the global equity market in the next two decades. The technology is also expected to drive the success of the AI chips market, with investments in AI processors by data centres increasing from $5bn a year to $22bn in 2025, the article detailed.

The report also identified 2020 as an innovative year for conversational AI. AI models, for example, could impersonate human-like understanding and language accuracy for the first time ever. Additionally, AI smart speakers responded to as many as 100 billion voice commands in 2020, a 75% increase from 2019. Experts further expect investments in conversational AI to grow in the near future, the article noted.

Username: Vala Afshar

Twitter handle: @ValaAfshar

Likes: 121

Retweets: 50

3. Dr Sally Eaves’ tweet on NIST’s call for inputs to develop AI risk management framework

Dr Sally Eaves, senior policy advisor at the Global Foundation for Cyber Studies and Research, a cybersecurity think tank, shared an article on the National Institute of Standards and Technology (NIST) issuing a call to the public to share inputs on developing a framework for managing AI risks. The Artificial Intelligence Risk Management Framework (AI RMF) will enable developers, users and evaluators to advance the reliability of AI models.

The call for inputs was made after the Congress and the White House requested the NIST, which is part of the US Department of Commerce, to develop such AI framework. Don Graves, Deputy Commerce Secretary, believes that the framework to manage AI risks to sensitive data and information will help in understanding and improving the capabilities of the US in the global competitive market for AI.

Username: Dr Sally Eaves

Twitter handle: @sallyeaves

Likes: 67

Retweets: 86

4. Mario Pawlowski’s tweet on Walmart testing its driverless delivery service

Mario Pawlowski, CEO of iTrucker, a logistics provider based in Illinois, US, shared an article on the retail company Walmart collaborating with software company Argo AI and automobile manufacturer Ford to launch driverless cars for delivering customers’ orders. The self-driving delivery service is being initially tested in three cities including Miami, Austin, and Washington DC.

Walmart is currently testing the driverless service with just a few autonomous vehicles, which deliver orders to the customers’ homes. Two trained drivers, however, are expected to be in the cars during the testing phase to ensure safety and reliability of the service. The retailer is also expected to expand the driverless service to other cities if the test proves successful.

Username: Mario Pawlowski

Twitter handle: @PawlowskiMario

Likes: 79

Retweets: 48

5. Dr Ganapathi Pulipaka’s tweet on the importance of data preparation

Dr Ganapathi Pulipaka, chief data scientist at Accenture, a professional services company dealing in cloud, digital, and security services, shared an article on why data preparation is crucial in AI workflows. AI workflows involve four steps including data preparation, modelling, simulation and testing, and deployment.

Some engineers emphasise that the modelling stage is most critical in delivering accurate information and insights, the article detailed. Other designers, however, believe that the data preparation stage is most important, as data is crucial to connecting all the phases in the AI workflows. Data preparation ensures that the right data or information is being included in the model, the article noted.

David Willingham, deep learning product manager at computer software company MathWorks, defines the first step in AI workflows to be the solution for any design problem. He added that engineers should not spend too much time on tuning an AI model, but rather focus on understanding the input data to ensure that the data is transformed into something meaningful.

Username: Dr Ganapathi Pulipaka

Twitter handle: @gp_pulipaka

Likes: 37

Retweets: 72