AI Bots out of work?
We all may be struggling to cope with how the world has changed but spare a thought for the AI bots. So called narrowly designed artificial intelligence have been helping organisations make decisions based on data from a pre-covid world. AI models are no longer capable of making predictions, suggesting a fundamental issue with how AI is being designed.
David Cox, IBM director of the MIT-IBM Watson AI Lab, explains that faulty AI is particularly problematic in the case of so-called black box predictive models: those algorithms which work in ways that are not visible, or understandable, to the user. “It’s very dangerous,” Cox says, “if you don’t understand what’s going on internally within a model in which you shovel data on one end to get a result on the other end. The model is supposed to embody the structure of the world, but there is no guarantee that it will keep working if the world changes.”
But still, spare a thought for Margaret, an AI bot trained to recognise patterns in speech. After being trained to analyse the unique speech pattern of President Donald Trump, the bot crashed. She has since been re-programmed to understand Trump’s unique syntax by analysing over 1 million words used by the President since 1976. Margaret is now able to decipher when President Trump is lying, when he is mad and how stressed he is when speaking on camera.
It’s not all bad for AI bots though. Call center workers have long been a frontier of workplace automation and in recent years, chatbots and voice assistants have made some headway. But with the onset of global work from home orders, the adoption of Contact Center AI has been accelerated. It’s not just reserved for the large multinationals. Google has launched their Rapid Response Virtual Agent, a special version of its Contact Agent AI. Many public sector organisations around the world have readily adopted the technology as they have been inundated with calls from citizens. Whilst the voice based agents that act like siri or Alexa may not be able to deal with complex calls, they have helped reduce call waiting times and filter to appropriate human handlers.
Hackers having a field day during Covid-19
Cyberattacks against banks have increased by 238% in the wake of the Coronavirus pandemic sweeping the globe. VMware Carbon Black released its third edition of the Modern Bank Heists report this week, which surveys top CIO’s including the top CISO’s from major financial institutions. In addition to wire fraud attempts, the report found that ransomware attacks increased 9x over the three months to end of April 2020.
“When combined with a steady commercial growth of mobile devices, cloud-based data storage and services, and digital payment systems, cybercriminals today have an ever-expanding host of attack vectors to exploit,” commented Jonah Force Hill, senior cyber policy advisor and CIAB executive director. “Every organization — providers of financial services, in particular — must remain vigilant in the face of these evolving threats. It is critical that organizations maintain a continuous dialogue with law enforcement to ensure a rapid response in the event of an incident.”
Cybercriminal activity has soared across the board in the past few months. Cybersecurity firm Bolster’s latest report into the “State of Phishing and Online Fraud Report” found 854,441 phishing and counterfeit pages and an additional 4 million suspicious pages for the first 3 months of the year.
Abhishek Dubey, its co-founder and CEO, sees a challenging future ahead.
“We anticipate phishing site creation will continue to increase, especially as we proceed further into a COVID-minded world. The phishing lures and tactics of cybercriminals will consistently evolve to keep up with the rapidly changing threat landscape, but the underlying credential theft will not,” Dubey said in a statement.
“Cybersecurity conscious organizations will need to work together and leverage AI, automation and security training to effectively combat phishing and online fraud during this surge and beyond,” he added.
It’s not just online that hackers are exploiting opportunities. This week researcher Björn Ruytenberg exposed a potential security flaw with Thunderbolt equipped PCs. The technique outlined by Ruytenburg allows potential hackers to take control of a machine via a security vulnerability in a so-called “evil maid” attack. The claim has been disputed by many manufacturers such as Dell and Intel.
Something for the weekend
Fancy docking at the international space station? Well now you can. Ahead of SpaceX’s first manned launch to the ISS at the end of the month, the talented engineers at SpaceX have developed a desktop simulator for anyone to test their space docking skills. (spoiler, it’s harder than it looks) SpaceX ISS Simulator