Written by John Gwinner
Today, unemployment is at a record low, yet according to Glassdoor, each corporate job offer attracts over 250 resumes. If a recruiter read these in depth, did internet searches to validate the resume, and scanned professional sites for connections, they could easily spend months looking for the right resume.
A human cannot handle this deluge. Cue the computer.
Artificial Intelligence. Everyone thinks it means something different. Many people think it means living, thinking computers. Like the old saying: “The amazing thing about a dancing bear isn’t that it dances well – it’s that it dances at all.”
These recruiting systems do use a primitive form of AI; typically, resume scanning for certain keywords. These have been the bane of job seekers everywhere; a friend was rejected because they didn’t have “Social Media” even though they liked their key skill, “Social Media Marketing.”
The bear does not dance well. The fact that it dances at all saves us from being buried, so everyone uses these systems, while at the same time despising them.
A few years ago, at the Game Developer’s Convention, at the AI roundtable, I remarked that I’d be happy if we had Artificial Stupidity. I mean, stupid would be if someone decided to offer me a CTO job in NYC if I had “cannot relocate” on my resume, but was otherwise a perfect fit (I live in SoCal).
At this point, I’d take that job. Sometimes, we can all use a little stupidity.
When I was in the Marine Corps, one of the “Jodys” that we called to keep in time running, had a stanza that I’d rarely heard anyone else call. The punchline of that stanza was “To show the world the brains we lack!”
It’s not smart to charge up a sandy beach with bullets whizzing past your ear. Just because you aren’t smart, doesn’t mean you’re stupid.
Newton isn’t smart to jump into the recruiting fracas. There are billions of resumes and jobs floating around. How do you sort through it all? They would have to be crazy.
Crazy like a fox.
How do they sort through it all? Via their new AI algorithms, they sort through it well.
Back to what most people think AI is: it sounds like an artificial intelligence; someone smart, that is fueled by electricity instead of Jolt Cola and Hostess Twinkies, like those developers under the desks keeping the economy going.
It’s not: it’s more like your toaster. Not hot enough? Heating filaments on. That’s it. AI requires very narrowly defined problems, usually expressible in a series of numbers. One of the popular AI training algorithms looks at flowers. They reduce the Iris down to four numbers; the width and length of the green parts and the pretty parts. These numbers get added up, and out pops a Virginia iris.
Computers are very fast at adding up these numbers perfectly. If we can reduce difficult problems to a series of numbers, computers can help give us the answer; it’s only as good as the developers behind it.
At the end of the day, it’s number crunching. Do you use Linear Regression or Gradient Descent? NLP (Natural Language Processing) can scan for meaning, not just keywords, but takes a long time if no words match. Where do we draw the line? You need someone pretty smart to figure out how to crunch those numbers.
Newton.AI has some very smart developers.
It’s all very empirical. You’ve got to spend the time with the data, going through it, running the analysis. Where do companies tuck away their job listings? Is it fresh or stale? Where is the right point to bring in NLP? That analysis is exhausting.
Yet, it started simply enough.
Helder Silva, CEO of Newton.ai, was a clinical researcher in a hospital pursuing his Ph.D. when he met his Co-founder, Rui Costa, at a Hackathon organized by Microsoft in Portugal. Rui, who at the time was finishing up his degree in Software Engineering, mentioned that “AI would be a gateway towards a fulfilling future.” They noticed that the youth unemployment rate in Europe was rising, with some countries reaching close to 45%; yet also there were substantial employment vacancies. There was a distribution problem and a mismatch between job seekers and companies, masked by the sheer quantities of resumes involved.
Now seeing the problem, Helder and Rui brainstormed solutions. They continue to refine their algorithm that reacts to the fluctuations of the market in real time and builds an efficient distribution model to balance demand and supply. The right candidates can be matched up with the right company, efficiently and effectively, saving time and money for all.
Newton.AI has done this exhaustive analysis. They don’t just rely on AI; they also personally review and hand curate every applicant. Their AI lets them concentrate on a smaller pool of higher quality experts.
As Helder says: “Recruitment will be completely reshaped by AI, and this will lead to a better and fair future for job recruitment, that is not biased in race, ethnicity, gender, age, or beliefs.”
Shouldn’t you put that to work for your recruiting needs? Isn’t it time for the bear to dance well, and quickly? Newton can deliver hand curated experts that are a fantastic fit within 48 hours of an opening. Can your recruiters do the same?
‘Spooky’ AI tool brings dead relatives’ photos to life
By Umberto Bacchi
(Thomson Reuters Foundation) – Like the animated paintings that adorn the walls of Harry Potter’s school, a new online tool promises to bring portraits of dead relatives to life, stirring debate about the use of technology to impersonate people.
Genealogy company MyHeritage launched its “Deep Nostalgia” feature earlier this week, allowing users to turn stills into short videos showing the person in the photograph smiling, winking and nodding.
“Seeing our beloved ancestors’ faces come to life … lets us imagine how they might have been in reality, and provides a profound new way of connecting to our family history,” MyHeritage founder Gilad Japhet said in a statement.
Developed with Israeli computer vision firm D-ID, Deep Nostalgia uses deep learning algorithms to animate images with facial expressions that were based on those of MyHeritage employees.
Some of the company’s users took to Twitter on Friday to share the animated images of their deceased relatives, as well as moving depictions of historical figures, including Albert Einstein and Ancient Egypt’s lost Queen Nefertiti.
“Takes my breath away. This is my grandfather who died when I was eight. @MyHeritage brought him back to life. Absolutely crazy,” wrote Twitter user Jenny Hawran.
While most expressed amazement, others described the feature as “spooky” and said it raised ethical questions. “The photos are enough. The dead have no say in this,” tweeted user Erica Cervini.
From chatbots to virtual reality, the tool is the latest innovation seeking to bring the dead to life through technology.
Last year U.S. rapper Kanye West famously gifted his wife Kim Kardashian a hologram of her late father congratulating her on her birthday and on marrying “the most, most, most, most, most genius man in the whole world”.
‘ANIMATING THE PAST’
The trend has opened up all sorts of ethical and legal questions, particularly around consent and the opportunity to blur reality by recreating a virtual doppelganger of the living.
Elaine Kasket a psychology professor at the University of Wolverhampton in Britain who authored a book on the “digital afterlife”, said that while Deep Nostalgia was not necessarily “problematic”, it sat “at the top of a slippery slope”.
“When people start overwriting history or sort of animating the past … You wonder where that ends up,” she said.
MyHeritage acknowledges on its website that the technology can be “a bit uncanny” and its use “controversial”, but said steps have been taken to prevent abuses.
“The Deep Nostalgia feature includes hard-coded animations that are intentionally without any speech and therefore cannot be used to fake any content or deliver any message,” MyHeritage public relations director Rafi Mendelsohn said in a statement.
Yet, images alone can convey meaning, said Faheem Hussain, a clinical assistant professor at Arizona State University’s School for the Future of Innovation in Society.
“Imagine somebody took a picture of the Last Supper and Judas is now winking at Mary Magdalene – what kind of implications that can have,” Hussain told the Thomson Reuters Foundation by phone.
Similarly, Artificial Intelligence (AI) animations could be use to make someone appear as though they were doing things they might not be happy about, such as rolling their eyes or smiling at a funeral, he added.
Mendelsohn of MyHeritage said using photos of a living person without their consent was a breach of the company’s terms and conditions, adding that videos were clearly marked with AI symbols to differentiate them from authentic recordings.
“It is our ethical responsibility to mark such synthetic videos clearly and differentiate them from real videos,” he said.
(Reporting by Umberto Bacchi @UmbertoBacchi in Milan; Editing by Helen Popper. Please credit the Thomson Reuters Foundation, the charitable arm of Thomson Reuters, that covers the lives of people around the world who struggle to live freely or fairly. Visit http://news.trust.org)
Does your institution have operational resilience? Testing cyber resilience may be a good way to find out
By Callum Roxan, Head of Threat Intelligence, F-Secure
If ever 2020 had a lesson, it was that no organization can possibly prepare for every conceivable outcome. Yet building one particular skill will make any crisis easier to handle: operational resilience.
Many financial institutions have already devoted resources to building operational resilience. Unfortunately, this often takes what Miles Celic, Chief Executive Officer of TheCityUK, calls a “near death” experience for this conversion to occur. “Recent years have seen a number of cases of loss of reputation, reduced enterprise value and senior executive casualties from operational incidents that have been badly handled,” he wrote.
But it need not take a disaster to learn this vital lesson.
“Operational resilience means not only planning around specific, identified risks,” Charlotte Gerken, the executive director of the Bank of England, said in a 2017 speech on operational resilience. “We want firms to plan on the assumption that any part of their infrastructure could be impacted, whatever the reason.” Gerken noted that firms that had successfully achieved a level of resilience that survives a crisis had established the necessary mechanisms to bring the business together to respond where and when risks materialised, no matter why or how.
We’ll talk about the bit we know best here; by testing for cyber resilience, a company can do more than prepare for the worst sort of attacks it may face. This process can help any business get a clearer view of how it operates, and how well it is prepared for all kinds of surprises.
Assumptions and the mechanisms they should produce are the best way to prepare for the unknown. But, as the boxer Mike Tyson once said, “Everyone has a plan until they get punched in the mouth.” The aim of cyber resilience is to build an effective security posture that survives that first punch, and the several that are likely to follow. So how can an institution be confident that they’ve achieved genuine operational resilience?
This requires an organization to honestly assess itself through the motto inscribed at the front of the Temple of Delphi: “Know thyself.” And when it comes to cyber security, there is a way for an organization to test just how thoroughly it comprehends its own strengths and weaknesses.
The Bank of England was the first central bank to help develop the framework for institutions to test the integrity of their systems. CBEST is made up of controlled, bespoke, intelligence-led cyber security tests that replicate behaviours of those threat actors, and often have unforeseen or secondary benefits. Gerken notes that the “firms that did best in the testing tended to be those that really understood their organisations. They understood their own needs, strengths and weaknesses, and reflected this in the way they built resilience.”
In short, testing cyber resilience can provide clear insight into an institution’s operational resilience in general.
Gaining that specific knowledge without a “near-death” experience is obviously a significant win for any establishment. And testing for operational resilience throughout the industry can provide some reminders of the steps every organization should take so that testing provides unique insists about their institution, and not just a checklist of cyber defence basics.
The IIF/McKinsey Cyber Resilience Survey of the financial services industry released in March lasy year provided six sets of immediate actions that institutions could take to improve their cyber security posture. The toplines of these recommendations were:
- Do the basics, patch your vulnerabilities.
- Review your cloud architecture and security capabilities.
- Reduce your supply chain risk.
- Practice your incident response and recovery capabilities.
- Set aside a specific cyber security budget and prioritise it
- Build a skilled talent pool and optimize resources through automation.
But let’s be honest: If simply reading a solid list of recommendations created cyber resilience, cyber criminals would be out of business. Unfortunately, cyber crime as a business is booming and threat actors targeting essential financial institutions through cyber attacks are likely earning billions in the trillion dollar industry of financial crime.A list can’t reveal an institution’s unique weaknesses, those security failings and chokepoints that could shudder operations, not just during a successful cyber attack but during various other crises that challenge their operations. And the failings that lead to flaws in an institution’s cyber defence likely reverberate throughout the organization as liabilities that other crises would likely expose.
The best way to get a sense of operational resilience will always be to simulate the worst that attackers can summon. That’s why the time to test yourself is now, before someone else does.
Thomson Reuters to stress AI, machine learning in a post-pandemic world
By Kenneth Li and Nick Zieminski
NEW YORK (Reuters) – Thomson Reuters Corp will streamline technology, close offices and rely more on machines to prepare for a post-pandemic world, the news and information group said on Tuesday, as it reported higher sales and operating profit.
The Toronto-headquartered company will spend $500 million to $600 million over two years to burnish its technology credentials, investing in AI and machine learning to get data faster to professional customers increasingly working from home during the coronavirus crisis.
It will transition from a content provider to a content-driven technology company, and from a holding company to an operational structure.
Thomson Reuters’ New York- and Toronto-listed shares each gained more than 8%.
It aims to cut annual operating expenses by $600 million through eliminating duplicate functions, modernizing and consolidating technology, as well as through attrition and shrinking its real estate footprint. Layoffs are not a focus of the cost cuts and there are no current plans to divest assets as part of this plan, the company said.
“We look at the changing behaviors as a result of COVID … on professionals working from home working remotely being much more reliant on 24-7, digital always-on, sort of real-time always available information, served through software and powered by AI and ML (machine learning),” Chief Executive Steve Hasker said in an interview.
Sales growth is forecast to accelerate in each of the next three years compared with 1.3% reported sales growth for 2020, the company said in its earnings release.
Thomson Reuters, which owns Reuters News, said revenues rose 2% to $1.62 billion, while its operating profit jumped more than 300% to $956 million, reflecting the sale of an investment and other items.
Its three main divisions, Legal Professionals, Tax & Accounting Professionals, and Corporates, all showed higher organic quarterly sales and adjusted profit. As part of the two-year change program, the corporate, legal and tax side will operate more as one customer-facing entity.
Adjusted earnings per share of 54 cents were ahead of the 46 cents expected, based on data from Refinitiv.
The company raised its annual dividend by 10 cents to $1.62 per share.
The Reuters News business showed lower revenue in the fourth quarter. In January, Stephen J. Adler, Reuters’ editor-in-chief for the past decade, said he would retire in April from the world’s largest international news provider.
Thomson Reuters also said its stake in The London Stock Exchange is now worth about $11.2 billion.
The LSE last month completed its $27-billion takeover of data and analytics business Refinitiv, 45%-owned by Thomson Reuters.
(Reporting by Ken Li, writing by Nick Zieminski in New York, editing by Louise Heavens and Jane Merriman)
Sunak to give UK Infrastructure Bank £12 billion of capital in budget
LONDON (Reuters) – British finance minister Rishi Sunak is expected to announce an initial 12 billion pounds of capital and...
Robinhood plans confidential IPO filing as soon as March – Bloomberg News
(Reuters) – Online brokerage Robinhood, at the centre of this year’s retail trading frenzy, is planning to file confidentially for...
Wall Street Week Ahead: Investors weigh new stock leadership as broader market wobbles
By Lewis Krauskopf NEW YORK (Reuters) – A shakeup in stocks accelerated by the past week’s surge in Treasury yields...
SoftBank reaches settlement with former WeWork CEO Neumann
(Reuters) – SoftBank Group Corp said on Friday it has reached a settlement with WeWork’s special committee and the company’s...
Sunak warns of bill to be paid to tackle Britain’s ‘exposed’ finances – FT
(Reuters) – British finance minister Rishi Sunak will use the budget next week to level with the public over the...